In a month I will be packing to go to India; this time for an extended
period. But its also a goodbye to Qi and computing. At some point
you have to acknowledge that Qi doesn't pay its way. It was fun
though and I'm not sad about it.
Qi has been a journey that began nearly 20 years ago when I was a very
different person and worked for a university. But the book on Qi II
marks a natural watershed. I need to move on. By September I will be
gone.
Its ironic that David found perhaps my essays and Ars Programma just
as I'm on my way out of the door and about to take the site down. The
essays are based on a talk given in Milan. You can read them here.
If I stay in computing I will inevitably be drawn into political and
educational issues around software and its use. I know how hard that
battle will be. It will use all my resources and final success is not
certain. I don't want to be banging on in the same way when I'm 60.
There has to a better life than that and I'm off to find it.
It's has been a pleasure to work with Qi, I can think of continue
using your software maybe improving on it, because I reallly like
your
effort, and enjoy working with it. Maybe it will not be a success and
maybe
Lisp people is not agreeing on your ideas, You see for me success of
Qi is not
the goal, It's to joyfully working with it in different ways trying to
implement
fun ideas and maybe boring details and maybe some day it will fly
maybe not,
How do you like Qi to continue?
And Mark, I have felt that you have been a bit alone in your struggle
here.
What I have missed is a more close cooperation on the future of Qi,
maybe
a chat forum to hang out while coding in Qi, and be able to quickly
have a chat
about some interesting topic.
Anyway, India will be a new adventure, and maybe new way of living
life
Have fun
Stefan
On 1 Aug, 09:49, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> In a month I will be packing to go to India; this time for an extended
> period. But its also a goodbye to Qi and computing. At some point
> you have to acknowledge that Qi doesn't pay its way. It was fun
> though and I'm not sad about it.
> Qi has been a journey that began nearly 20 years ago when I was a very
> different person and worked for a university. But the book on Qi II
> marks a natural watershed. I need to move on. By September I will be
> gone.
> Its ironic that David found perhaps my essays and Ars Programma just
> as I'm on my way out of the door and about to take the site down. The
> essays are based on a talk given in Milan. You can read them here.
> If I stay in computing I will inevitably be drawn into political and
> educational issues around software and its use. I know how hard that
> battle will be. It will use all my resources and final success is not
> certain. I don't want to be banging on in the same way when I'm 60.
> There has to a better life than that and I'm off to find it.
On 1 Aug, 11:01, snorgers <stefan.ta...@spray.se> wrote:
> Thanks for your all your effort Mark!,
> It's has been a pleasure to work with Qi,
Thankyou.
I can think of continue
> using your software maybe improving on it, because I reallly like
> your
> effort, and enjoy working with it. Maybe it will not be a success and
> maybe
> Lisp people is not agreeing on your ideas, You see for me success of
> Qi is not
> the goal, It's to joyfully working with it in different ways trying to
> implement
> fun ideas and maybe boring details and maybe some day it will fly
> maybe not,
Thats good to know.
> How do you like Qi to continue?
I think that the general thrust is contained in my essay on the next
Lisp.
Find what is really primitive in Qi and what you need to make it run.
Start by flattening the Lisp source and grabbing the set of
capitalised functions. See what subset is really primitive. From
there you have a base to move Qi to different platforms by mapping the
kernel. CL seems to be going nowhere. Clojure or Python or Ruby
would be good.
If Qi is detached from CL then it will no longer be taken for granted
how the user calls the OS, invokes streams etc. So you will need a
standard. Look at CLTL and consider what exactly is *really* needed.
Aim to be as spartan as possible because being spartan will make
porting easy.
Aside from that, you need to build up the library because libraries
make or break languages. Remember if you follow my suggestion about
kernelisation and making Qi 'viral' over the entire Lisp genotype
(i.e. Lispy languages in general and I include Python in that) then
basing your libraries on CL features outside the kernel is not a good
way to go. You will maroon your work in CL which is dying. Write your
libraries as far as possible in Qi itself.
> And Mark, I have felt that you have been a bit alone in your struggle
> here.
> What I have missed is a more close cooperation on the future of Qi,
> maybe
> a chat forum to hang out while coding in Qi, and be able to quickly
> have a chat
> about some interesting topic.
Well here is a place to do that. There are nearly 200 people here.
Thats why I created the group (also I know bugger all about making
message boards and it was the easiest option). So its a space for
people to use. I don't censor anything, I just clean off the spam.
> Anyway, India will be a new adventure, and maybe new way of living
> life
I hope so. Perhaps on my return I'll see things in a rosier light.
Mark, what exactly do you mean with "taking the site down"? I've just
begun studying Qi and would hate to see the whole things go away.
I can contribute with funds to cover the site upkeep costs, or
alternatively make a local backup of the site and upload it to Google
Code or something similiar. Which would you prefer?
Cheers,
Babar
On 1 Aug, 14:39, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> On 1 Aug, 11:01, snorgers <stefan.ta...@spray.se> wrote:
> > Thanks for your all your effort Mark!,
> > It's has been a pleasure to work with Qi,
> Thankyou.
> I can think of continue
> > using your software maybe improving on it, because I reallly like
> > your
> > effort, and enjoy working with it. Maybe it will not be a success and
> > maybe
> > Lisp people is not agreeing on your ideas, You see for me success of
> > Qi is not
> > the goal, It's to joyfully working with it in different ways trying to
> > implement
> > fun ideas and maybe boring details and maybe some day it will fly
> > maybe not,
> Thats good to know.
> > How do you like Qi to continue?
> I think that the general thrust is contained in my essay on the next
> Lisp.
> Find what is really primitive in Qi and what you need to make it run.
> Start by flattening the Lisp source and grabbing the set of
> capitalised functions. See what subset is really primitive. From
> there you have a base to move Qi to different platforms by mapping the
> kernel. CL seems to be going nowhere. Clojure or Python or Ruby
> would be good.
> If Qi is detached from CL then it will no longer be taken for granted
> how the user calls the OS, invokes streams etc. So you will need a
> standard. Look at CLTL and consider what exactly is *really* needed.
> Aim to be as spartan as possible because being spartan will make
> porting easy.
> Aside from that, you need to build up the library because libraries
> make or break languages. Remember if you follow my suggestion about
> kernelisation and making Qi 'viral' over the entire Lisp genotype
> (i.e. Lispy languages in general and I include Python in that) then
> basing your libraries on CL features outside the kernel is not a good
> way to go. You will maroon your work in CL which is dying. Write your
> libraries as far as possible in Qi itself.
> > And Mark, I have felt that you have been a bit alone in your struggle
> > here.
> > What I have missed is a more close cooperation on the future of Qi,
> > maybe
> > a chat forum to hang out while coding in Qi, and be able to quickly
> > have a chat
> > about some interesting topic.
> Well here is a place to do that. There are nearly 200 people here.
> Thats why I created the group (also I know bugger all about making
> message boards and it was the easiest option). So its a space for
> people to use. I don't censor anything, I just clean off the spam.
> > Anyway, India will be a new adventure, and maybe new way of living
> > life
> I hope so. Perhaps on my return I'll see things in a rosier light.
I agree with this. What is meant by "taking the site down"? I would
very much like to see the Qi site continue. If the site must come
down, then putting the content into Google Code would be a good way to
preserve it. I have limited funds, but I could help with site upkeep
in a small way.
Also, if possible, it would be good to have a discussion about the
"kernel of Qi" and how to carry it forward.
The Qi type system looks very interesting. Having an implementation
of that type system in Clojure and Python strikes me as a very useful
thing.
On 1 Aug, 09:27, "Babar K. Zafar" <babar.za...@gmail.com> wrote:
> Mark, what exactly do you mean with "taking the site down"? I've just
> begun studying Qi and would hate to see the whole things go away.
> I can contribute with funds to cover the site upkeep costs, or
> alternatively make a local backup of the site and upload it to Google
> Code or something similiar. Which would you prefer?
> Cheers,
> Babar
> On 1 Aug, 14:39, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > On 1 Aug, 11:01, snorgers <stefan.ta...@spray.se> wrote:
> > > Thanks for your all your effort Mark!,
> > > It's has been a pleasure to work with Qi,
> > Thankyou.
> > I can think of continue
> > > using your software maybe improving on it, because I reallly like
> > > your
> > > effort, and enjoy working with it. Maybe it will not be a success and
> > > maybe
> > > Lisp people is not agreeing on your ideas, You see for me success of
> > > Qi is not
> > > the goal, It's to joyfully working with it in different ways trying to
> > > implement
> > > fun ideas and maybe boring details and maybe some day it will fly
> > > maybe not,
> > Thats good to know.
> > > How do you like Qi to continue?
> > I think that the general thrust is contained in my essay on the next
> > Lisp.
> > Find what is really primitive in Qi and what you need to make it run.
> > Start by flattening the Lisp source and grabbing the set of
> > capitalised functions. See what subset is really primitive. From
> > there you have a base to move Qi to different platforms by mapping the
> > kernel. CL seems to be going nowhere. Clojure or Python or Ruby
> > would be good.
> > If Qi is detached from CL then it will no longer be taken for granted
> > how the user calls the OS, invokes streams etc. So you will need a
> > standard. Look at CLTL and consider what exactly is *really* needed.
> > Aim to be as spartan as possible because being spartan will make
> > porting easy.
> > Aside from that, you need to build up the library because libraries
> > make or break languages. Remember if you follow my suggestion about
> > kernelisation and making Qi 'viral' over the entire Lisp genotype
> > (i.e. Lispy languages in general and I include Python in that) then
> > basing your libraries on CL features outside the kernel is not a good
> > way to go. You will maroon your work in CL which is dying. Write your
> > libraries as far as possible in Qi itself.
> > > And Mark, I have felt that you have been a bit alone in your struggle
> > > here.
> > > What I have missed is a more close cooperation on the future of Qi,
> > > maybe
> > > a chat forum to hang out while coding in Qi, and be able to quickly
> > > have a chat
> > > about some interesting topic.
> > Well here is a place to do that. There are nearly 200 people here.
> > Thats why I created the group (also I know bugger all about making
> > message boards and it was the easiest option). So its a space for
> > people to use. I don't censor anything, I just clean off the spam.
> > > Anyway, India will be a new adventure, and maybe new way of living
> > > life
> > I hope so. Perhaps on my return I'll see things in a rosier light.
> Mark, what exactly do you mean with "taking the site down"? I've just
> begun studying Qi and would hate to see the whole things go away.
> I can contribute with funds to cover the site upkeep costs, or
> alternatively make a local backup of the site and upload it to Google
> Code or something similiar. Which would you prefer?
> Cheers,
> Babar
> On 1 Aug, 14:39, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > On 1 Aug, 11:01, snorgers <stefan.ta...@spray.se> wrote:
> > > Thanks for your all your effort Mark!,
> > > It's has been a pleasure to work with Qi,
> > Thankyou.
> > I can think of continue
> > > using your software maybe improving on it, because I reallly like
> > > your
> > > effort, and enjoy working with it. Maybe it will not be a success and
> > > maybe
> > > Lisp people is not agreeing on your ideas, You see for me success of
> > > Qi is not
> > > the goal, It's to joyfully working with it in different ways trying to
> > > implement
> > > fun ideas and maybe boring details and maybe some day it will fly
> > > maybe not,
> > Thats good to know.
> > > How do you like Qi to continue?
> > I think that the general thrust is contained in my essay on the next
> > Lisp.
> > Find what is really primitive in Qi and what you need to make it run.
> > Start by flattening the Lisp source and grabbing the set of
> > capitalised functions. See what subset is really primitive. From
> > there you have a base to move Qi to different platforms by mapping the
> > kernel. CL seems to be going nowhere. Clojure or Python or Ruby
> > would be good.
> > If Qi is detached from CL then it will no longer be taken for granted
> > how the user calls the OS, invokes streams etc. So you will need a
> > standard. Look at CLTL and consider what exactly is *really* needed.
> > Aim to be as spartan as possible because being spartan will make
> > porting easy.
> > Aside from that, you need to build up the library because libraries
> > make or break languages. Remember if you follow my suggestion about
> > kernelisation and making Qi 'viral' over the entire Lisp genotype
> > (i.e. Lispy languages in general and I include Python in that) then
> > basing your libraries on CL features outside the kernel is not a good
> > way to go. You will maroon your work in CL which is dying. Write your
> > libraries as far as possible in Qi itself.
> > > And Mark, I have felt that you have been a bit alone in your struggle
> > > here.
> > > What I have missed is a more close cooperation on the future of Qi,
> > > maybe
> > > a chat forum to hang out while coding in Qi, and be able to quickly
> > > have a chat
> > > about some interesting topic.
> > Well here is a place to do that. There are nearly 200 people here.
> > Thats why I created the group (also I know bugger all about making
> > message boards and it was the easiest option). So its a space for
> > people to use. I don't censor anything, I just clean off the spam.
> > > Anyway, India will be a new adventure, and maybe new way of living
> > > life
> > I hope so. Perhaps on my return I'll see things in a rosier light.
I don't want to bring this question up, but reading your notice above
the question of what
happens to the Qi if the book goes out in print or if Qi is risking
this at some time in the future
got it's attention. Actually I was a little in chock after the first
read of your notice.
I don't want to ask you to release it as a BSD license or similar. But
is there a possibility that you can
connect the license to a more time-resistable payment rule such as
"either a purchase of the book, or a receipt that you
put x amount of cash on a bank account" or something similar. Then
people will still probably buy the book and if it goes out of
print there still be an option to get a license by putting in money on
a bank account.
Maybe this is a no-problem or this is not the solution, but I think
it's good to have this discussion to make things clear.
Regards
Stefan
On 1 Aug, 19:05, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> I'm not taking Lambda Associates down. Just Ars Programma.
> Mark
> On 1 Aug, 14:27, "Babar K. Zafar" <babar.za...@gmail.com> wrote:
> > Mark, what exactly do you mean with "taking the site down"? I've just
> > begun studying Qi and would hate to see the whole things go away.
> > I can contribute with funds to cover the site upkeep costs, or
> > alternatively make a local backup of the site and upload it to Google
> > Code or something similiar. Which would you prefer?
> > Cheers,
> > Babar
> > On 1 Aug, 14:39, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > > On 1 Aug, 11:01, snorgers <stefan.ta...@spray.se> wrote:
> > > > Thanks for your all your effort Mark!,
> > > > It's has been a pleasure to work with Qi,
> > > Thankyou.
> > > I can think of continue
> > > > using your software maybe improving on it, because I reallly like
> > > > your
> > > > effort, and enjoy working with it. Maybe it will not be a success and
> > > > maybe
> > > > Lisp people is not agreeing on your ideas, You see for me success of
> > > > Qi is not
> > > > the goal, It's to joyfully working with it in different ways trying to
> > > > implement
> > > > fun ideas and maybe boring details and maybe some day it will fly
> > > > maybe not,
> > > Thats good to know.
> > > > How do you like Qi to continue?
> > > I think that the general thrust is contained in my essay on the next
> > > Lisp.
> > > Find what is really primitive in Qi and what you need to make it run.
> > > Start by flattening the Lisp source and grabbing the set of
> > > capitalised functions. See what subset is really primitive. From
> > > there you have a base to move Qi to different platforms by mapping the
> > > kernel. CL seems to be going nowhere. Clojure or Python or Ruby
> > > would be good.
> > > If Qi is detached from CL then it will no longer be taken for granted
> > > how the user calls the OS, invokes streams etc. So you will need a
> > > standard. Look at CLTL and consider what exactly is *really* needed.
> > > Aim to be as spartan as possible because being spartan will make
> > > porting easy.
> > > Aside from that, you need to build up the library because libraries
> > > make or break languages. Remember if you follow my suggestion about
> > > kernelisation and making Qi 'viral' over the entire Lisp genotype
> > > (i.e. Lispy languages in general and I include Python in that) then
> > > basing your libraries on CL features outside the kernel is not a good
> > > way to go. You will maroon your work in CL which is dying. Write your
> > > libraries as far as possible in Qi itself.
> > > > And Mark, I have felt that you have been a bit alone in your struggle
> > > > here.
> > > > What I have missed is a more close cooperation on the future of Qi,
> > > > maybe
> > > > a chat forum to hang out while coding in Qi, and be able to quickly
> > > > have a chat
> > > > about some interesting topic.
> > > Well here is a place to do that. There are nearly 200 people here.
> > > Thats why I created the group (also I know bugger all about making
> > > message boards and it was the easiest option). So its a space for
> > > people to use. I don't censor anything, I just clean off the spam.
> > > > Anyway, India will be a new adventure, and maybe new way of living
> > > > life
> > > I hope so. Perhaps on my return I'll see things in a rosier light.
> Actually I was a little in chock after the first
> read of your notice.
Don't take it too hard. I've actually had something like this in mind
for a long time now. Perhaps the sudden announcement was startling,
but I'm more practised now in recognising when situations need to be
changed and moving on and I move quickly. People tend to hang on too
long to old situations that don't reward them. I'd like to earn some
money and it would be fun doing something new.
Growth is a vector quantity and not a scalar one. I've stopped
growing along the computer science axis and instead I'll grow along
another axis. So Mark Tarver, computer scientist dies to be replaced
by Mark Tarver, philosopher, poet, therapist, ditch digger, whatever.
Its only when you stop growing on every axis that you're stuck. There
are some ostensibly successful people - lawyers, accountants, GPL
advocates ( ;) ) who are stuck on every axis.
Generally I've found that my interests are cyclic. So maybe in a
future cycle I'll return to CS.
Regarding your work on kernelisation; what you've taken on here, is no
small affair. People would love you to put Qi on Clojure/Python etc.
but 95% of them will be content to watch you work your balls off
implementing it. They won't help you. At the end of your labour they
will step forward to claim the rewards and they will want it free as
in free beer and no strings (BSD/MIT). *It will be your baby so you
will decide for yourself.* If it blows a fuse you will be expected to
fix it and that will be free beer and no strings too. Your motives
are noble, but many of those in FOSS are not.
I have no panacea for making human beings better than they are.
Perhaps the best advice I can offer is in my book on Taoism; fourth
discourse,
Fu Hsiang: The unenlightened superior man believes in the existence of
good and evil. He tries to do good and make a better world. Yet the
unpredictable nature of the world means his designs are often
frustrated. Friends let him down; money meant for a good cause is
embezzled and squandered. Corruption gets in the way of the noblest
project. When the superior man encounters these things, he is dismayed
in his heart. He feels his failure keenly and blames those around him.
In this way, the noblest natures are apt to fall into bitterness when
confronted with the reality of the world and the schemes of inferior
men.
In contrast, the superior man who is on a conscious journey to become
a sage, begins his journey with the understanding that good and evil
are illusory and that his good acts are only a means to an end. He
proceeds by acting like a sage, doing right, but separating himself
from the fruit of his action. He recognises correctly that the results
of his actions are of less importance than the intent behind them.
Since he does not trouble himself with the events of the world, he
avoids bitterness and pain.
> I don't want to ask you to release it as a BSD license or similar. But
> is there a possibility that you can
> connect the license to a more time-resistable payment rule such as
> "either a purchase of the book, or a receipt that you
> put x amount of cash on a bank account" or something similar. Then
> people will still probably buy the book and if it goes out of
> print there still be an option to get a license by putting in money on
> a bank account.
> Maybe this is a no-problem or this is not the solution, but I think
> it's good to have this discussion to make things clear.
> Regards
> Stefan
> On 1 Aug, 19:05, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > I'm not taking Lambda Associates down. Just Ars Programma.
> > Mark
> > On 1 Aug, 14:27, "Babar K. Zafar" <babar.za...@gmail.com> wrote:
> > > Mark, what exactly do you mean with "taking the site down"? I've just
> > > begun studying Qi and would hate to see the whole things go away.
> > > I can contribute with funds to cover the site upkeep costs, or
> > > alternatively make a local backup of the site and upload it to Google
> > > Code or something similiar. Which would you prefer?
> > > Cheers,
> > > Babar
> > > On 1 Aug, 14:39, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > > > > It's has been a pleasure to work with Qi,
> > > > Thankyou.
> > > > I can think of continue
> > > > > using your software maybe improving on it, because I reallly like
> > > > > your
> > > > > effort, and enjoy working with it. Maybe it will not be a success and
> > > > > maybe
> > > > > Lisp people is not agreeing on your ideas, You see for me success of
> > > > > Qi is not
> > > > > the goal, It's to joyfully working with it in different ways trying to
> > > > > implement
> > > > > fun ideas and maybe boring details and maybe some day it will fly
> > > > > maybe not,
> > > > Thats good to know.
> > > > > How do you like Qi to continue?
> > > > I think that the general thrust is contained in my essay on the next
> > > > Lisp.
> > > > Find what is really primitive in Qi and what you need to make it run.
> > > > Start by flattening the Lisp source and grabbing the set of
> > > > capitalised functions. See what subset is really primitive. From
> > > > there you have a base to move Qi to different platforms by mapping the
> > > > kernel. CL seems to be going nowhere. Clojure or Python or Ruby
> > > > would be good.
> > > > If Qi is detached from CL then it will no longer be taken for granted
> > > > how the user calls the OS, invokes streams etc. So you will need a
> > > > standard. Look at CLTL and consider what exactly is *really* needed.
> > > > Aim to be as spartan as possible because being spartan will make
> > > > porting easy.
> > > > Aside from that, you need to build up the library because libraries
> > > > make or break languages. Remember if you follow my suggestion about
> > > > kernelisation and making Qi 'viral' over the entire Lisp genotype
> > > > (i.e. Lispy languages in general and I include Python in that) then
> > > > basing your libraries on CL features outside the kernel is not a good
> > > > way to go. You will maroon your work in CL which is dying. Write your
> > > > libraries as far as possible in Qi itself.
> > > > > And Mark, I have felt that you have been a bit alone in your struggle
> > > > > here.
> > > > > What I have missed is a more close cooperation on the future of Qi,
> > > > > maybe
> > > > > a chat forum to hang out while coding in Qi, and be able to quickly
> > > > > have a chat
> > > > > about some interesting topic.
> > > > Well here is a place to do that. There are nearly 200 people here.
> > > > Thats why I created the group (also I know bugger all about making
> > > > message boards and it was the easiest option). So its a space for
> > > > people to use. I don't censor anything, I just clean off the spam.
> > > > > Anyway, India will be a new adventure, and maybe new way of living
> > > > > life
> > > > I hope so. Perhaps on my return I'll see things in a rosier light.
> Regarding your work on kernelisation; what you've taken on here, is no
> small affair. People would love you to put Qi on Clojure/Python etc.
> but 95% of them will be content to watch you work your balls off
> implementing it. They won't help you. At the end of your labour they
> will step forward to claim the rewards and they will want it free as
> in free beer and no strings (BSD/MIT). *It will be your baby so you
> will decide for yourself.* If it blows a fuse you will be expected to
> fix it and that will be free beer and no strings too. Your motives
> are noble, but many of those in FOSS are not.
I don't mind paying for a book or pay you directly. Actually from
having
gone through all your Qi work, I think I prefere to pay you something
like buying
the book which I did. Anyway I don't mind taking up the ball and try
to reformulate qi in a more
kernelized way. I guess I'm just weired. But I will do it on my own
pace if I
have to do this alone. Also I will code it in a fun and artistic and
fast way, that is
the port will be dead slow at start. Code can always be rewritten
later. I will try to
attract people to coloborate with me because I'm a great person to
know and work
with / learn from, just that I will be more lonely if there is
uncertainty of the license.
But what the heck I'm pretty tough, I can manage. And yes I will
release code under BSD/GNU/MIT
whatever because coding is a social thing for me, I really don't care
for money. I'm pretty well trained
and does a decent job, giving an ok salary. Actually coding in Qi is a
great mental educator.
So if I have to do this alone, It _will_ take considerable time.
If I do things for free, and a fuse go off, I will fix it, but not
promptly
whatever they say.
True, there is a point and meaning whatever we do in life.
But, it's the journey that makes life worth living.
(Freely translated from the Swedish "word-gardener" Karin Boye)
Regards
Stefan
On 4 Aug, 20:24, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > Actually I was a little in chock after the first
> > read of your notice.
> Don't take it too hard. I've actually had something like this in mind
> for a long time now. Perhaps the sudden announcement was startling,
> but I'm more practised now in recognising when situations need to be
> changed and moving on and I move quickly. People tend to hang on too
> long to old situations that don't reward them. I'd like to earn some
> money and it would be fun doing something new.
> Growth is a vector quantity and not a scalar one. I've stopped
> growing along the computer science axis and instead I'll grow along
> another axis. So Mark Tarver, computer scientist dies to be replaced
> by Mark Tarver, philosopher, poet, therapist, ditch digger, whatever.
> Its only when you stop growing on every axis that you're stuck. There
> are some ostensibly successful people - lawyers, accountants, GPL
> advocates ( ;) ) who are stuck on every axis.
> Generally I've found that my interests are cyclic. So maybe in a
> future cycle I'll return to CS.
> Regarding your work on kernelisation; what you've taken on here, is no
> small affair. People would love you to put Qi on Clojure/Python etc.
> but 95% of them will be content to watch you work your balls off
> implementing it. They won't help you. At the end of your labour they
> will step forward to claim the rewards and they will want it free as
> in free beer and no strings (BSD/MIT). *It will be your baby so you
> will decide for yourself.* If it blows a fuse you will be expected to
> fix it and that will be free beer and no strings too. Your motives
> are noble, but many of those in FOSS are not.
> I have no panacea for making human beings better than they are.
> Perhaps the best advice I can offer is in my book on Taoism; fourth
> discourse,
> Fu Hsiang: The unenlightened superior man believes in the existence of
> good and evil. He tries to do good and make a better world. Yet the
> unpredictable nature of the world means his designs are often
> frustrated. Friends let him down; money meant for a good cause is
> embezzled and squandered. Corruption gets in the way of the noblest
> project. When the superior man encounters these things, he is dismayed
> in his heart. He feels his failure keenly and blames those around him.
> In this way, the noblest natures are apt to fall into bitterness when
> confronted with the reality of the world and the schemes of inferior
> men.
> In contrast, the superior man who is on a conscious journey to become
> a sage, begins his journey with the understanding that good and evil
> are illusory and that his good acts are only a means to an end. He
> proceeds by acting like a sage, doing right, but separating himself
> from the fruit of his action. He recognises correctly that the results
> of his actions are of less importance than the intent behind them.
> Since he does not trouble himself with the events of the world, he
> avoids bitterness and pain.
> Mark
> > I don't want to ask you to release it as a BSD license or similar. But
> > is there a possibility that you can
> > connect the license to a more time-resistable payment rule such as
> > "either a purchase of the book, or a receipt that you
> > put x amount of cash on a bank account" or something similar. Then
> > people will still probably buy the book and if it goes out of
> > print there still be an option to get a license by putting in money on
> > a bank account.
> > Maybe this is a no-problem or this is not the solution, but I think
> > it's good to have this discussion to make things clear.
> > Regards
> > Stefan
> > On 1 Aug, 19:05, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > > I'm not taking Lambda Associates down. Just Ars Programma.
> > > Mark
> > > On 1 Aug, 14:27, "Babar K. Zafar" <babar.za...@gmail.com> wrote:
> > > > Mark, what exactly do you mean with "taking the site down"? I've just
> > > > begun studying Qi and would hate to see the whole things go away.
> > > > I can contribute with funds to cover the site upkeep costs, or
> > > > alternatively make a local backup of the site and upload it to Google
> > > > Code or something similiar. Which would you prefer?
> > > > Cheers,
> > > > Babar
> > > > On 1 Aug, 14:39, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > > > > > Thanks for your all your effort Mark!,
> > > > > > It's has been a pleasure to work with Qi,
> > > > > Thankyou.
> > > > > I can think of continue
> > > > > > using your software maybe improving on it, because I reallly like
> > > > > > your
> > > > > > effort, and enjoy working with it. Maybe it will not be a success and
> > > > > > maybe
> > > > > > Lisp people is not agreeing on your ideas, You see for me success of
> > > > > > Qi is not
> > > > > > the goal, It's to joyfully working with it in different ways trying to
> > > > > > implement
> > > > > > fun ideas and maybe boring details and maybe some day it will fly
> > > > > > maybe not,
> > > > > Thats good to know.
> > > > > > How do you like Qi to continue?
> > > > > I think that the general thrust is contained in my essay on the next
> > > > > Lisp.
> > > > > Find what is really primitive in Qi and what you need to make it run.
> > > > > Start by flattening the Lisp source and grabbing the set of
> > > > > capitalised functions. See what subset is really primitive. From
> > > > > there you have a base to move Qi to different platforms by mapping the
> > > > > kernel. CL seems to be going nowhere. Clojure or Python or Ruby
> > > > > would be good.
> > > > > If Qi is detached from CL then it will no longer be taken for granted
> > > > > how the user calls the OS, invokes streams etc. So you will need a
> > > > > standard. Look at CLTL and consider what exactly is *really* needed.
> > > > > Aim to be as spartan as possible because being spartan will make
> > > > > porting easy.
> > > > > Aside from that, you need to build up the library because libraries
> > > > > make or break languages. Remember if you follow my suggestion about
> > > > > kernelisation and making Qi 'viral' over the entire Lisp genotype
> > > > > (i.e. Lispy languages in general and I include Python in that) then
> > > > > basing your libraries on CL features outside the kernel is not a good
> > > > > way to go. You will maroon your work in CL which is dying. Write your
> > > > > libraries as far as possible in Qi itself.
> > > > > > And Mark, I have felt that you have been a bit alone in your struggle
> > > > > > here.
> > > > > > What I have missed is a more close cooperation on the future of Qi,
> > > > > > maybe
> > > > > > a chat forum to hang out while coding in Qi, and be able to quickly
> > > > > > have a chat
> > > > > > about some interesting topic.
> > > > > Well here is a place to do that. There are nearly 200 people here.
> > > > > Thats why I created the group (also I know bugger all about making
> > > > > message boards and it was the easiest option). So its a space for
> > > > > people to use. I don't censor anything, I just clean off the spam.
> > > > > > Anyway, India will be a new adventure, and maybe new way of living
> > > > > > life
> > > > > I hope so. Perhaps on my return I'll see things in a rosier light.
If you bought the book that's enough. That was the deal. The book
was only really a token. The money is unimportant to me as such - it
is such a small sum. Simply a token of respect nothing more. Its
always important to give something back when one receives. It is as
much for the character of the recipient as for the giver to do this.
So you are free to do as you want. You have no obligation.
Certainly you should do it at your own pace and in the any way you see
fit. And yes, try to attract people to help you. It is not a small
task. It is better for the work and for people's hearts that many
contribute instead of one. Then it is a dedication.
Seeing Stefan's explorations of Qi and his involvement in it certainly
has been and remains an inspiration to me. And this, although I've
been mostly silent here in the last several months. (And never did
anything more here than conversing with you.) I may sometime start
offering my help. As of now, though, I'm more concentrated on
improving my career path with clojure/java. When I'm happy with the
direction my career is taking, I might help.
Anyway, this post is much more about Stefan than about my never-coming
help. I just wanted to say I've seen enough of Stefan to know I, for
one, would be happy to someday collaborate and follow his leadership
in being proactive about Qi.
Thanks Mark for bringing Qi to us. There's no month that passes in
which I don't think to myself how I would like someday to be in a
position where I can integrate it into my daily toys; and contribute
back to you.
On Aug 4, 5:30 pm, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> If you bought the book that's enough. That was the deal. The book
> was only really a token. The money is unimportant to me as such - it
> is such a small sum. Simply a token of respect nothing more. Its
> always important to give something back when one receives. It is as
> much for the character of the recipient as for the giver to do this.
> So you are free to do as you want. You have no obligation.
> Certainly you should do it at your own pace and in the any way you see
> fit. And yes, try to attract people to help you. It is not a small
> task. It is better for the work and for people's hearts that many
> contribute instead of one. Then it is a dedication.
I wasn't aware that Stefan's first target for Shen is Clojure. I'm
definitely pleased with this choice. Moreover, although our soon-to-be
family (my wife is halfway through her pregnancy) is in the process of
buying its first home, I *do* have much more free time recently than I
would usually; and I expect this situation to last for at least a few
more months. Thus, I decided I'd like to help Stefan starting today. I
don't need to spend all my time on Clojure/java alone.
So I'll dust off my Qi book and start a second reading. It'll be much
faster to read than it was the first time. While I'm doing this, I
could help porting easy stuff, slowly ramping up on the difficulty.
Although I'm a quick learner, I believe I'll need some help and lots
of code reviews.
All that I ask is that you may think of the easiest things I could
start helping you with.
Sincerely,
Daniel
On Aug 26, 12:14 pm, Daniel Jomphe <danieljom...@gmail.com> wrote:
> Seeing Stefan's explorations of Qi and his involvement in it certainly
> has been and remains an inspiration to me. And this, although I've
> been mostly silent here in the last several months. (And never did
> anything more here than conversing with you.) I may sometime start
> offering my help. As of now, though, I'm more concentrated on
> improving my career path with clojure/java. When I'm happy with the
> direction my career is taking, I might help.
> Anyway, this post is much more about Stefan than about my never-coming
> help. I just wanted to say I've seen enough of Stefan to know I, for
> one, would be happy to someday collaborate and follow his leadership
> in being proactive about Qi.
> Thanks Mark for bringing Qi to us. There's no month that passes in
> which I don't think to myself how I would like someday to be in a
> position where I can integrate it into my daily toys; and contribute
> back to you.
> On Aug 4, 5:30 pm, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > If you bought the book that's enough. That was the deal. The book
> > was only really a token. The money is unimportant to me as such - it
> > is such a small sum. Simply a token of respect nothing more. Its
> > always important to give something back when one receives. It is as
> > much for the character of the recipient as for the giver to do this.
> > So you are free to do as you want. You have no obligation.
> > Certainly you should do it at your own pace and in the any way you see
> > fit. And yes, try to attract people to help you. It is not a small
> > task. It is better for the work and for people's hearts that many
> > contribute instead of one. Then it is a dedication.
> I wasn't aware that Stefan's first target for Shen is Clojure. I'm
> definitely pleased with this choice. Moreover, although our soon-to-be
> family (my wife is halfway through her pregnancy) is in the process of
> buying its first home, I *do* have much more free time recently than I
> would usually; and I expect this situation to last for at least a few
> more months. Thus, I decided I'd like to help Stefan starting today. I
> don't need to spend all my time on Clojure/java alone.
> So I'll dust off my Qi book and start a second reading. It'll be much
> faster to read than it was the first time. While I'm doing this, I
> could help porting easy stuff, slowly ramping up on the difficulty.
> Although I'm a quick learner, I believe I'll need some help and lots
> of code reviews.
> All that I ask is that you may think of the easiest things I could
> start helping you with.
> Sincerely,
> Daniel
> On Aug 26, 12:14 pm, Daniel Jomphe <danieljom...@gmail.com> wrote:
> > Seeing Stefan's explorations of Qi and his involvement in it certainly
> > has been and remains an inspiration to me. And this, although I've
> > been mostly silent here in the last several months. (And never did
> > anything more here than conversing with you.) I may sometime start
> > offering my help. As of now, though, I'm more concentrated on
> > improving my career path with clojure/java. When I'm happy with the
> > direction my career is taking, I might help.
> > Anyway, this post is much more about Stefan than about my never-coming
> > help. I just wanted to say I've seen enough of Stefan to know I, for
> > one, would be happy to someday collaborate and follow his leadership
> > in being proactive about Qi.
> > Thanks Mark for bringing Qi to us. There's no month that passes in
> > which I don't think to myself how I would like someday to be in a
> > position where I can integrate it into my daily toys; and contribute
> > back to you.
> > On Aug 4, 5:30 pm, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > > If you bought the book that's enough. That was the deal. The book
> > > was only really a token. The money is unimportant to me as such - it
> > > is such a small sum. Simply a token of respect nothing more. Its
> > > always important to give something back when one receives. It is as
> > > much for the character of the recipient as for the giver to do this.
> > > So you are free to do as you want. You have no obligation.
> > > Certainly you should do it at your own pace and in the any way you see
> > > fit. And yes, try to attract people to help you. It is not a small
> > > task. It is better for the work and for people's hearts that many
> > > contribute instead of one. Then it is a dedication.
> I don't want to bring this question up, but reading your notice above
> the question of what
> happens to the Qi if the book goes out in print or if Qi is risking
> this at some time in the future
> got it's attention. Actually I was a little in chock after the first
> read of your notice.
> I don't want to ask you to release it as a BSD license or similar. But
> is there a possibility that you can
> connect the license to a more time-resistable payment rule such as
> "either a purchase of the book, or a receipt that you
> put x amount of cash on a bank account" or something similar. Then
> people will still probably buy the book and if it goes out of
> print there still be an option to get a license by putting in money on
> a bank account.
> Maybe this is a no-problem or this is not the solution, but I think
> it's good to have this discussion to make things clear.
OK here is the thread retitled to a discussion.
*************************************************************************** ******
Inevitably Qi is going to move to a more liberal license and it should
match Shen. So that might answer your query.
*************************************************************************** *******
However I wanted to say this. What follows is a bit rambling.
The problems with license and financing etc. that have come up again
and again have proved hard to resolve. I used to think that that was
just my ignorance, but the more I have read into this, the more I
realise that we have a generic problem with revenue models for
software, music, newspapers and perhaps in the future, books too. I
also used to think that freeloaders of all kinds had some very clever
general economic model which I had not grasped. I kept waiting to
'get it' and never have. I've since come to the conclusion that they
do not have a general model and they simply BS.
Here is an interesting article and a very long and interesting thread
on 'freeloading' (free downloading) music.
It's worth reading because the commentators include people in the
music business and my view is contained in the comment that is that
'there is no argument so feeble or lacking in conviction that people
will not employ it to take the things they want.' Some of the
freeloading arguments read like a satire on Mr Bumble (e.g. keeping
musicians poor helps them make good music) or how about 'copyright
should not exist because songs use English and that is free'. In
that thread you will find nearly all the arguments about FOSS
replicated in music. But what stands out is the immaturity and
recklessness of the freeloaders and their refusal to extend respect to
those that point out quite reasonably, that economically the Emperor
has no clothes.
The refusal of freeloaders to engage with the economic side coupled
with the prevalence of poor arguments buttressed by bad manners
doesn't go down with people like me who want answers. See
Having some idiot berate you for not making your work free because he
conceives he has a right to it, just sets back their cause. ** In
fact, I would say that Qi would have appeared by now under BSD/MIT but
for the awful online performance of the FOSSers; being called a
'moron' doesn't really move me to be liberal to them.** It certainly
doesn't move me to want to labour for free in their service.
What I can see happening is that GPL version 3 etc. is not going to
stick. Young geeks think Stallman belongs to yesterday. Stallman has
not changed, but online users have and they want FREE (BSD/MIT) not
'free speech' not GPL; see
However the economic problem is not going to go away and the central
economic question is still to be answered. There is one viable model
of development as I stated in my essay.
"Good software arises when one or more very good programmers work
closely full time together over a period of time developing,
maintaining and improving it."
We need solutions of how to sustain this model if we abandon
conventional marketing. People cannot live on air.
Solutions?
==========
My view is that what we have here is a social problem and we need
social solutions to this. I think that partly we need to enforce
copyright; especially for musicians, otherwise the poor buggers will
be homeless. And Pirate Bay needs to be closed down. But still there
is a problem. We need to make FOSS work.
I think there are two general solutions.
The first is to open to FOSSers the funding that is at present
earmarked purely for unis (e.g. 750m for IT in Framework 7).
The second is to create an open source educational framework like EBay
where people can exchange money for tuition from geeks who know their
stuff. That was the idea of Ars Programma. However it needs
cooperation from more than one person. Clearly the working model
that gave rise to Qi is not it.
QUOTE
but I think it's good to have this discussion to make things clear.
UNQUOTE
Btw I wanted to say that since Stefan wanted a discussion; I'm just
responding to him and making this space in the last few days I am
here. I'm not preempting any conclusion you want to reach. It's
simply a space for you to debate (or not) his and your preferences in
whatever way you want. If everything is clear and satisfactory, then
there is no need to debate or take this thread further. But if there
is something that you want to debate or resolve for which you feel you
need me, then now is the time.
My interest and the effort I put into Qi and in the future, Shen, is for intellectual stimulation rather than the prospect of any financial gain or remuneration. I think Qi offers a very interesting framework for the study of computer language structure and development and we are indebted to you for providing this software playground. My feeling is that we should encourage people to join this project or at least play with the result by making the licence terms as permissive as possible so my vote is for both Qi and Shen to be released under the modified BSD or MIT but many of the alternative FOSS licences in:
would probably also be fine by me if people have a particular preference. For those who dislike the possibility/prospect that their effort on Qi/Shen might be included in a commercial product and they would like to disallow this then a FSF GPL or equivalent would seem suitable and also fine by me. My company releases OpenFOAM under the GPL to avoid our work being ripped-off and we sell support and consultancy services to pay for our work and this is an effective strategy for this product. However, it is not clear if there would be an interest in including Qi/Shen in a commercial product and even if there might be it is not clear that this should be discouraged and that is why I would be happy for any of my work on Qi or Shen to be released under a more permissive licence like modified BSD or MIT if you think that this is most appropriate for the project as a whole.
Generally GPL 2 proved not to be popular for Qi I. I don't think we
can go back there.
GPL 3 will be no more popular. You could effectively simply give to
all people the freedom given by the commercial licence which comes
with the book which simply keeps the sources open and allows you to do
with your work as you want.
I guess that BSD would be widely understood and as of now seems the
logical choice.
Qi was designed to be programmable using sugar functions and be shaped
like putty to look like whatever you want. This means that Qi can
spawn a whole host of experimental languages mounted on different
platforms. In philosophy this is the diametric opposite of Python
(fixed syntax and format). A sort of creative chaos can ensue and I'm
not too unhappy about that. I think people need to be free to pursue
their ideas - I certainly did. And inevitably other people's ideas
are going to differ from mine. It would be good though to have some
sort of core for people to come back to - some sort of standard which
is guaranteed to run. A kind of drum beat to hold together some
frenzied jamming sessions.
Generally I think that the way to do this is to be cautious in hacking
the sources too much, to be liberal in building in features that
*expect* that people will want to do their own thing. Sugar functions,
programmable comment characters are examples and this approach can be
extended. We want to avoid a situation where there are umpteen
experimental copies each of which require people to reinstall Qi/Shen
just to run experimental ideas. We want to be able to run these ideas
by simply loading them into the standard.
I'm going to give some thought to this aspect and esp. that of
allowing people to optimise the generated code for different platforms
(Clojure, CL) using the type information gleaned by type checking.
The standard needs to accomodate the powerful ideas of viral infection
outlined in 'The Next Lisp'.
Mark
On 30 Aug, 10:25, Henry Weller <HWell...@googlemail.com> wrote:
> My interest and the effort I put into Qi and in the future, Shen, is for
> intellectual stimulation rather than the prospect of any financial gain or
> remuneration. I think Qi offers a very interesting framework for the study of
> computer language structure and development and we are indebted to you for
> providing this software playground. My feeling is that we should encourage
> people to join this project or at least play with the result by making the
> licence terms as permissive as possible so my vote is for both Qi and Shen to be
> released under the modified BSD or MIT but many of the alternative FOSS licences
> in:
> would probably also be fine by me if people have a particular preference. For
> those who dislike the possibility/prospect that their effort on Qi/Shen might be
> included in a commercial product and they would like to disallow this then a FSF
> GPL or equivalent would seem suitable and also fine by me. My company releases
> OpenFOAM under the GPL to avoid our work being ripped-off and we sell support
> and consultancy services to pay for our work and this is an effective strategy
> for this product. However, it is not clear if there would be an interest in
> including Qi/Shen in a commercial product and even if there might be it is not
> clear that this should be discouraged and that is why I would be happy for any
> of my work on Qi or Shen to be released under a more permissive licence like
> modified BSD or MIT if you think that this is most appropriate for the project
> as a whole.
I then suggest that the Shen code beeing released under the BSD
license.
There is one thing to note, I take Qi code and modify it. Therefor
Mark has to
bless the new code in order to be able to release under a BSD
license.
I will continue to rework the Qi base, introduce only crusial features
that
can only with difficulty be implemented in library code or with sugar
functions
e.g. patching the core code or there is an architectual nicity to do
it in
the core.
I have other thing that I'm planning to implement, one is the ability
to memoize
the ptolog process. E.g. you record your progv:s and then replay them
under
a condition of the underlying code is not stateful and that the
arguments is considered
the same as the previously done deduction. By using similar ideas as
git has to track
information, the check that something is similar to a previous case
can optionally use
a key instead of comparing a whole structure. This is actually a risk,
but people may wan't
to buy into that risk if there is a significant speed improvement
compared to the failure risk.
Why it is so nice to use hash values is that you can easilly define
sameness of a tree with
another tree if it is in a ekvivalent class of one or many commutating
rules,
E.G Consider updating the hash with
Key <- Key(a) + Key(b) mod p if a and b commutes
Key <- Key(a) + Mixer * Key(b) mod p if a and b does not commute
I have worked out the overall strategy to do this but First I will
continue to hack on core.qi
/Stefan
On 30 Aug, 13:54, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> Generally GPL 2 proved not to be popular for Qi I. I don't think we
> can go back there.
> GPL 3 will be no more popular. You could effectively simply give to
> all people the freedom given by the commercial licence which comes
> with the book which simply keeps the sources open and allows you to do
> with your work as you want.
> I guess that BSD would be widely understood and as of now seems the
> logical choice.
> Qi was designed to be programmable using sugar functions and be shaped
> like putty to look like whatever you want. This means that Qi can
> spawn a whole host of experimental languages mounted on different
> platforms. In philosophy this is the diametric opposite of Python
> (fixed syntax and format). A sort of creative chaos can ensue and I'm
> not too unhappy about that. I think people need to be free to pursue
> their ideas - I certainly did. And inevitably other people's ideas
> are going to differ from mine. It would be good though to have some
> sort of core for people to come back to - some sort of standard which
> is guaranteed to run. A kind of drum beat to hold together some
> frenzied jamming sessions.
> Generally I think that the way to do this is to be cautious in hacking
> the sources too much, to be liberal in building in features that
> *expect* that people will want to do their own thing. Sugar functions,
> programmable comment characters are examples and this approach can be
> extended. We want to avoid a situation where there are umpteen
> experimental copies each of which require people to reinstall Qi/Shen
> just to run experimental ideas. We want to be able to run these ideas
> by simply loading them into the standard.
> I'm going to give some thought to this aspect and esp. that of
> allowing people to optimise the generated code for different platforms
> (Clojure, CL) using the type information gleaned by type checking.
> The standard needs to accomodate the powerful ideas of viral infection
> outlined in 'The Next Lisp'.
> Mark
> On 30 Aug, 10:25, Henry Weller <HWell...@googlemail.com> wrote:
> > Hi Mark,
> > My interest and the effort I put into Qi and in the future, Shen, is for
> > intellectual stimulation rather than the prospect of any financial gain or
> > remuneration. I think Qi offers a very interesting framework for the study of
> > computer language structure and development and we are indebted to you for
> > providing this software playground. My feeling is that we should encourage
> > people to join this project or at least play with the result by making the
> > licence terms as permissive as possible so my vote is for both Qi and Shen to be
> > released under the modified BSD or MIT but many of the alternative FOSS licences
> > in:
> > would probably also be fine by me if people have a particular preference. For
> > those who dislike the possibility/prospect that their effort on Qi/Shen might be
> > included in a commercial product and they would like to disallow this then a FSF
> > GPL or equivalent would seem suitable and also fine by me. My company releases
> > OpenFOAM under the GPL to avoid our work being ripped-off and we sell support
> > and consultancy services to pay for our work and this is an effective strategy
> > for this product. However, it is not clear if there would be an interest in
> > including Qi/Shen in a commercial product and even if there might be it is not
> > clear that this should be discouraged and that is why I would be happy for any
> > of my work on Qi or Shen to be released under a more permissive licence like
> > modified BSD or MIT if you think that this is most appropriate for the project
> > as a whole.
> Generally GPL 2 proved not to be popular for Qi I. I don't think we > can go back there. > GPL 3 will be no more popular. You could effectively simply give to > all people the freedom given by the commercial licence which comes > with the book which simply keeps the sources open and allows you to do > with your work as you want.
My understanding is that the GPL does not preclude commercial use but it does preclude inclusion in a closed-source commercial product. Whereas the FPQi licence precludes use for a commercial purpose unless the book is purchased which then also allows inclusion in a closed-source commercial product. It seems to me that the aims and motivations of the two approaches are rather different.
> I guess that BSD would be widely understood and as of now seems the > logical choice.
I agree, it makes the licencing very simple and easily understood to developers, contributors and users.
> Qi was designed to be programmable using sugar functions and be shaped > like putty to look like whatever you want. This means that Qi can > spawn a whole host of experimental languages mounted on different > platforms. In philosophy this is the diametric opposite of Python > (fixed syntax and format). A sort of creative chaos can ensue and I'm > not too unhappy about that.
I agree, and it is partly this freedom and flexibility that drew we in. What I have found though is that I cannot achieve my aims simply via sugar functions and that hacking the core to some extent has been necessary. This in turn has required a significant reorganisation of the code to support a boot-strap build system so that each version can be built using the previous version and generate the pure Lisp along the way so that it can be distributed without the complete build chain. I hope this smooth build system will be useful in the development for Shen and perhaps some of the ideas I am testing in Qi could also be accepted into Shen. Currently most of my work inherits your FPQi licence as required and the bits which are independent of your code are GPL but as yet unreleased. I would be happy to release everything I have done and will do on Qi/Shen under the modified BSD licence if there is an interest in this work.
> I think people need to be free to pursue > their ideas - I certainly did. And inevitably other people's ideas > are going to differ from mine. It would be good though to have some > sort of core for people to come back to - some sort of standard which > is guaranteed to run. A kind of drum beat to hold together some > frenzied jamming sessions.
I agree, but at this stage I think people need to play with the core a bit and then see what comes out of this and put together the best ideas for the ultimately stable core.
> Generally I think that the way to do this is to be cautious in hacking > the sources too much, to be liberal in building in features that > *expect* that people will want to do their own thing. Sugar functions, > programmable comment characters are examples and this approach can be > extended.
I am not convinced about programmable comment characters because variation at this level makes code portability, comprehension and the programming of font-locking and automated indentation into editors difficult.
> We want to avoid a situation where there are umpteen > experimental copies each of which require people to reinstall Qi/Shen > just to run experimental ideas. We want to be able to run these ideas > by simply loading them into the standard.
I agree, but is now the time to define this fixed core of functionality? Or is the definition still evolving? And if so would it be a good idea if experimental copies are created as a testing ground for the pieces of the eventual stable core?
Methodologically I cannot insist that the development proceeds
according to a set of dicta. *However contrariwise, people can't
impose their dicta on me.* Hence if a day comes where "Functional
Programming in Qi" becomes "Functional Programming in Shen", the new
stuff in the language must be stuff of which I approve. I won't
endorse what I don't like. And I won't be around to lay down the law
anyway even if I wanted to. So I'll judge when I return after a few
months and see what has been created.
Its too early to say what people will evolve. I can't bless what I
have not seen. I'm not going to limit experiment. But generally there
is a methodology here which I think is good.
1. Features unless well documented, highly reliable, stable and
clearly useful in a wide range of applications should be in a
library. I'm generous about libraries but stingy about core. CL made
a mistake here which Shen should not repeat. Stuff that looks
experimental goes there. After probation it may go into the core.
That includes O-O; ThinCL is in a library.
2. Experiment is best conducted *insofar as possible* in the
programmable part of Qi. This programmable part I can increase. Its
helpful to everyone if stuff can be run without hacking core.
3. Porting is a seperate issue from language standard and mixing the
two together is not a good idea.
Stuff you know I can approve of and bless for core:
a. Programmable pattern matching (I've got most of the code for that).
b. Threads
c. Exception handling*
d. Parallel processing
e. OS communication*
f. Foreign language interface
g. porting to Python or Clojure
* In Qilib
That's quite a list.
What Stefan is doing is interesting but I can't make a judgement until
I see what comes out of it. Its only been a few days since he
began.
Mark
18:48, Henry Weller <HWell...@googlemail.com> wrote:
> > Generally GPL 2 proved not to be popular for Qi I. I don't think we
> > can go back there.
> > GPL 3 will be no more popular. You could effectively simply give to
> > all people the freedom given by the commercial licence which comes
> > with the book which simply keeps the sources open and allows you to do
> > with your work as you want.
> My understanding is that the GPL does not preclude commercial use but it does
> preclude inclusion in a closed-source commercial product. Whereas the FPQi
> licence precludes use for a commercial purpose unless the book is purchased
> which then also allows inclusion in a closed-source commercial product. It
> seems to me that the aims and motivations of the two approaches are rather
> different.
> > I guess that BSD would be widely understood and as of now seems the
> > logical choice.
> I agree, it makes the licencing very simple and easily understood to developers,
> contributors and users.
> > Qi was designed to be programmable using sugar functions and be shaped
> > like putty to look like whatever you want. This means that Qi can
> > spawn a whole host of experimental languages mounted on different
> > platforms. In philosophy this is the diametric opposite of Python
> > (fixed syntax and format). A sort of creative chaos can ensue and I'm
> > not too unhappy about that.
> I agree, and it is partly this freedom and flexibility that drew we in. What I
> have found though is that I cannot achieve my aims simply via sugar functions
> and that hacking the core to some extent has been necessary. This in turn has
> required a significant reorganisation of the code to support a boot-strap build
> system so that each version can be built using the previous version and generate
> the pure Lisp along the way so that it can be distributed without the complete
> build chain. I hope this smooth build system will be useful in the development
> for Shen and perhaps some of the ideas I am testing in Qi could also be accepted
> into Shen. Currently most of my work inherits your FPQi licence as required
> and the bits which are independent of your code are GPL but as yet unreleased.
> I would be happy to release everything I have done and will do on Qi/Shen under
> the modified BSD licence if there is an interest in this work.
> > I think people need to be free to pursue
> > their ideas - I certainly did. And inevitably other people's ideas
> > are going to differ from mine. It would be good though to have some
> > sort of core for people to come back to - some sort of standard which
> > is guaranteed to run. A kind of drum beat to hold together some
> > frenzied jamming sessions.
> I agree, but at this stage I think people need to play with the core a bit and
> then see what comes out of this and put together the best ideas for the
> ultimately stable core.
> > Generally I think that the way to do this is to be cautious in hacking
> > the sources too much, to be liberal in building in features that
> > *expect* that people will want to do their own thing. Sugar functions,
> > programmable comment characters are examples and this approach can be
> > extended.
> I am not convinced about programmable comment characters because variation at
> this level makes code portability, comprehension and the programming of
> font-locking and automated indentation into editors difficult.
> > We want to avoid a situation where there are umpteen
> > experimental copies each of which require people to reinstall Qi/Shen
> > just to run experimental ideas. We want to be able to run these ideas
> > by simply loading them into the standard.
> I agree, but is now the time to define this fixed core of functionality? Or is
> the definition still evolving? And if so would it be a good idea if
> experimental copies are created as a testing ground for the pieces of the
> eventual stable core?
> What Stefan is doing is interesting but I can't make a judgement until
> I see what comes out of it. Its only been a few days since he
> began.
Pretty reasonable statement.
i'll put up some code for you to read about reduce_help, this is
needed so that
we understand where I'm heading and can combine code-bases and ideas.
Let me just comment
it a little more.
Also I think that I would like to use Henry's work. Apperently you can
generate a good
part of qi from qi-sources to lisp using QiII and his build system. So
this is a nice little
feature. I never gotten that to work myself, just some specific
functions. Is that ok with you
Henry!
> Also I think that I would like to use Henry's work. Apperently you can > generate a good > part of qi from qi-sources to lisp using QiII and his build system. So > this is a nice little > feature. I never gotten that to work myself, just some specific > functions. Is that ok with you > Henry!
I am glad that you have found my efforts useful and I will be taking this further still by generating more of the Lisp from Qi source. I am more than happy for you to use whatever I have done with Qi in whatever way suits your purpose for Qi or Shen and would be happy to change the licence on my parts from GPL to the modified BSD if Mark chooses that way to go for Qi and Shen.
> Its too early to say what people will evolve. I can't bless what I > have not seen. I'm not going to limit experiment. But generally there > is a methodology here which I think is good.
I would be happy to release everything I have done with Qi so far and in the future for you and other to review, use, discard, as you see fit, but under what licence?
** As a footnote I would say that porting Qi to Clojure/Python is more
immediately important and conceptually more straightforward than
changing the standard. Shen as I conceived it is 95% Qi. **
What Stefan is doing is trying to unify three strands of Qi into one.
Interestingly in the '80s a lot of people did similar work and it
never caught on. This is not to say it was wrong; that's just the way
it is. One can speculate why it never caught on and perhaps it was
because the computational models that evolved were too complex for
programmers to feel comfortable with. People aren't always smart
enough to be able to use multi-paradigm models.
Mark
On 30 Aug, 20:02, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> Methodologically I cannot insist that the development proceeds
> according to a set of dicta. *However contrariwise, people can't
> impose their dicta on me.* Hence if a day comes where "Functional
> Programming in Qi" becomes "Functional Programming in Shen", the new
> stuff in the language must be stuff of which I approve. I won't
> endorse what I don't like. And I won't be around to lay down the law
> anyway even if I wanted to. So I'll judge when I return after a few
> months and see what has been created.
> Its too early to say what people will evolve. I can't bless what I
> have not seen. I'm not going to limit experiment. But generally there
> is a methodology here which I think is good.
> 1. Features unless well documented, highly reliable, stable and
> clearly useful in a wide range of applications should be in a
> library. I'm generous about libraries but stingy about core. CL made
> a mistake here which Shen should not repeat. Stuff that looks
> experimental goes there. After probation it may go into the core.
> That includes O-O; ThinCL is in a library.
> 2. Experiment is best conducted *insofar as possible* in the
> programmable part of Qi. This programmable part I can increase. Its
> helpful to everyone if stuff can be run without hacking core.
> 3. Porting is a seperate issue from language standard and mixing the
> two together is not a good idea.
> Stuff you know I can approve of and bless for core:
> a. Programmable pattern matching (I've got most of the code for that).
> b. Threads
> c. Exception handling*
> d. Parallel processing
> e. OS communication*
> f. Foreign language interface
> g. porting to Python or Clojure
> * In Qilib
> That's quite a list.
> What Stefan is doing is interesting but I can't make a judgement until
> I see what comes out of it. Its only been a few days since he
> began.
> Mark
> 18:48, Henry Weller <HWell...@googlemail.com> wrote:
> > > Generally GPL 2 proved not to be popular for Qi I. I don't think we
> > > can go back there.
> > > GPL 3 will be no more popular. You could effectively simply give to
> > > all people the freedom given by the commercial licence which comes
> > > with the book which simply keeps the sources open and allows you to do
> > > with your work as you want.
> > My understanding is that the GPL does not preclude commercial use but it does
> > preclude inclusion in a closed-source commercial product. Whereas the FPQi
> > licence precludes use for a commercial purpose unless the book is purchased
> > which then also allows inclusion in a closed-source commercial product. It
> > seems to me that the aims and motivations of the two approaches are rather
> > different.
> > > I guess that BSD would be widely understood and as of now seems the
> > > logical choice.
> > I agree, it makes the licencing very simple and easily understood to developers,
> > contributors and users.
> > > Qi was designed to be programmable using sugar functions and be shaped
> > > like putty to look like whatever you want. This means that Qi can
> > > spawn a whole host of experimental languages mounted on different
> > > platforms. In philosophy this is the diametric opposite of Python
> > > (fixed syntax and format). A sort of creative chaos can ensue and I'm
> > > not too unhappy about that.
> > I agree, and it is partly this freedom and flexibility that drew we in. What I
> > have found though is that I cannot achieve my aims simply via sugar functions
> > and that hacking the core to some extent has been necessary. This in turn has
> > required a significant reorganisation of the code to support a boot-strap build
> > system so that each version can be built using the previous version and generate
> > the pure Lisp along the way so that it can be distributed without the complete
> > build chain. I hope this smooth build system will be useful in the development
> > for Shen and perhaps some of the ideas I am testing in Qi could also be accepted
> > into Shen. Currently most of my work inherits your FPQi licence as required
> > and the bits which are independent of your code are GPL but as yet unreleased.
> > I would be happy to release everything I have done and will do on Qi/Shen under
> > the modified BSD licence if there is an interest in this work.
> > > I think people need to be free to pursue
> > > their ideas - I certainly did. And inevitably other people's ideas
> > > are going to differ from mine. It would be good though to have some
> > > sort of core for people to come back to - some sort of standard which
> > > is guaranteed to run. A kind of drum beat to hold together some
> > > frenzied jamming sessions.
> > I agree, but at this stage I think people need to play with the core a bit and
> > then see what comes out of this and put together the best ideas for the
> > ultimately stable core.
> > > Generally I think that the way to do this is to be cautious in hacking
> > > the sources too much, to be liberal in building in features that
> > > *expect* that people will want to do their own thing. Sugar functions,
> > > programmable comment characters are examples and this approach can be
> > > extended.
> > I am not convinced about programmable comment characters because variation at
> > this level makes code portability, comprehension and the programming of
> > font-locking and automated indentation into editors difficult.
> > > We want to avoid a situation where there are umpteen
> > > experimental copies each of which require people to reinstall Qi/Shen
> > > just to run experimental ideas. We want to be able to run these ideas
> > > by simply loading them into the standard.
> > I agree, but is now the time to define this fixed core of functionality? Or is
> > the definition still evolving? And if so would it be a good idea if
> > experimental copies are created as a testing ground for the pieces of the
> > eventual stable core?
Just see it as a way to have a tool for building tools. In qi-prolog
you do matching, which you should not need to redo because this have
been done already in reduce-help. This means that the effort of
introducing
new features is done only ones. Say that you would like to be able to
match
segments in the sequent analysis, in qi-prolog and in a define and in
qi-yacc.
Now all you need is to overwrite the list stream object with the
append features in
and off you go, say that you would like to be able to use an array in
qi-yacc and
in qi-prolog for some odd reason - you then make a array object and
register it
as a stream and off you go, it should work everywhere. Say that you
would like to
make new accesors, write a small library, and register it acording to
a protocal, bam
you can use it everywhere. This is my vision for this rewrite. It's
just to powerful for
not to be done.
/Stefan
On 30 Aug, 22:51, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> ** As a footnote I would say that porting Qi to Clojure/Python is more
> immediately important and conceptually more straightforward than
> changing the standard. Shen as I conceived it is 95% Qi. **
> What Stefan is doing is trying to unify three strands of Qi into one.
> Interestingly in the '80s a lot of people did similar work and it
> never caught on. This is not to say it was wrong; that's just the way
> it is. One can speculate why it never caught on and perhaps it was
> because the computational models that evolved were too complex for
> programmers to feel comfortable with. People aren't always smart
> enough to be able to use multi-paradigm models.
> Mark
> On 30 Aug, 20:02, Mark Tarver <dr.mtar...@ukonline.co.uk> wrote:
> > Methodologically I cannot insist that the development proceeds
> > according to a set of dicta. *However contrariwise, people can't
> > impose their dicta on me.* Hence if a day comes where "Functional
> > Programming in Qi" becomes "Functional Programming in Shen", the new
> > stuff in the language must be stuff of which I approve. I won't
> > endorse what I don't like. And I won't be around to lay down the law
> > anyway even if I wanted to. So I'll judge when I return after a few
> > months and see what has been created.
> > Its too early to say what people will evolve. I can't bless what I
> > have not seen. I'm not going to limit experiment. But generally there
> > is a methodology here which I think is good.
> > 1. Features unless well documented, highly reliable, stable and
> > clearly useful in a wide range of applications should be in a
> > library. I'm generous about libraries but stingy about core. CL made
> > a mistake here which Shen should not repeat. Stuff that looks
> > experimental goes there. After probation it may go into the core.
> > That includes O-O; ThinCL is in a library.
> > 2. Experiment is best conducted *insofar as possible* in the
> > programmable part of Qi. This programmable part I can increase. Its
> > helpful to everyone if stuff can be run without hacking core.
> > 3. Porting is a seperate issue from language standard and mixing the
> > two together is not a good idea.
> > Stuff you know I can approve of and bless for core:
> > a. Programmable pattern matching (I've got most of the code for that).
> > b. Threads
> > c. Exception handling*
> > d. Parallel processing
> > e. OS communication*
> > f. Foreign language interface
> > g. porting to Python or Clojure
> > * In Qilib
> > That's quite a list.
> > What Stefan is doing is interesting but I can't make a judgement until
> > I see what comes out of it. Its only been a few days since he
> > began.
> > Mark
> > 18:48, Henry Weller <HWell...@googlemail.com> wrote:
> > > > Generally GPL 2 proved not to be popular for Qi I. I don't think we
> > > > can go back there.
> > > > GPL 3 will be no more popular. You could effectively simply give to
> > > > all people the freedom given by the commercial licence which comes
> > > > with the book which simply keeps the sources open and allows you to do
> > > > with your work as you want.
> > > My understanding is that the GPL does not preclude commercial use but it does
> > > preclude inclusion in a closed-source commercial product. Whereas the FPQi
> > > licence precludes use for a commercial purpose unless the book is purchased
> > > which then also allows inclusion in a closed-source commercial product. It
> > > seems to me that the aims and motivations of the two approaches are rather
> > > different.
> > > > I guess that BSD would be widely understood and as of now seems the
> > > > logical choice.
> > > I agree, it makes the licencing very simple and easily understood to developers,
> > > contributors and users.
> > > > Qi was designed to be programmable using sugar functions and be shaped
> > > > like putty to look like whatever you want. This means that Qi can
> > > > spawn a whole host of experimental languages mounted on different
> > > > platforms. In philosophy this is the diametric opposite of Python
> > > > (fixed syntax and format). A sort of creative chaos can ensue and I'm
> > > > not too unhappy about that.
> > > I agree, and it is partly this freedom and flexibility that drew we in. What I
> > > have found though is that I cannot achieve my aims simply via sugar functions
> > > and that hacking the core to some extent has been necessary. This in turn has
> > > required a significant reorganisation of the code to support a boot-strap build
> > > system so that each version can be built using the previous version and generate
> > > the pure Lisp along the way so that it can be distributed without the complete
> > > build chain. I hope this smooth build system will be useful in the development
> > > for Shen and perhaps some of the ideas I am testing in Qi could also be accepted
> > > into Shen. Currently most of my work inherits your FPQi licence as required
> > > and the bits which are independent of your code are GPL but as yet unreleased.
> > > I would be happy to release everything I have done and will do on Qi/Shen under
> > > the modified BSD licence if there is an interest in this work.
> > > > I think people need to be free to pursue
> > > > their ideas - I certainly did. And inevitably other people's ideas
> > > > are going to differ from mine. It would be good though to have some
> > > > sort of core for people to come back to - some sort of standard which
> > > > is guaranteed to run. A kind of drum beat to hold together some
> > > > frenzied jamming sessions.
> > > I agree, but at this stage I think people need to play with the core a bit and
> > > then see what comes out of this and put together the best ideas for the
> > > ultimately stable core.
> > > > Generally I think that the way to do this is to be cautious in hacking
> > > > the sources too much, to be liberal in building in features that
> > > > *expect* that people will want to do their own thing. Sugar functions,
> > > > programmable comment characters are examples and this approach can be
> > > > extended.
> > > I am not convinced about programmable comment characters because variation at
> > > this level makes code portability, comprehension and the programming of
> > > font-locking and automated indentation into editors difficult.
> > > > We want to avoid a situation where there are umpteen
> > > > experimental copies each of which require people to reinstall Qi/Shen
> > > > just to run experimental ideas. We want to be able to run these ideas
> > > > by simply loading them into the standard.
> > > I agree, but is now the time to define this fixed core of functionality? Or is
> > > the definition still evolving? And if so would it be a good idea if
> > > experimental copies are created as a testing ground for the pieces of the
> > > eventual stable core?
|
