Re: guile user base

Re: guile user base

[Anton Vidovic]

Hello,

I'm not subscribed to this ML, so please forgive me for breaking your
thread.

> In your opinion, how does guile user community compare (in size)
> with other free schemes user communities?

Compared by the number of installations alone, guile is probably the
most widely installed Scheme at all.

Contributing to that is the fact that it is a core dependency of the
game Aisle Riot, which is part of Gnome Games, which has been
installed by default on every Ubuntu installation for the last several
years, and probably by other Gnome centric distributions too.

* http://packages.debian.org/sid/aisleriot

* http://qa.debian.org/popcon-graph.php?packages=racket+chicken-bin+guile-1.8+&show_installed=on&want_legend=on&from_date=&to_date=&hlght_date=&date_fmt=%25Y-%25m&beenhere=1

More general, guile is probably also the most widely installed Lisp,
dwarfed in the number of installations maybe only by Emacs+Elisp.

> I am under the impression that guile user base is somewhat smaller
> than chicken' or racket', but to what extent?

One indicator could be the number of third party modules. The number
of chicken eggs

* http://wiki.call-cc.org/chicken-projects/egg-index-4.html

and racket PLaneT

* http://planet.racket-lang.org/

easily shadow the number of guile projects

* http://www.gnu.org/software/guile/gnu-guile-projects.html

> And how could we roughly figure this out? Also, how well is the user
> base growing since v2 came out? Do we have some download stats for
> instance?

I think, without checking anything, that the average number of users
on #guile has at least doubled since 2.0, going from roughly ~15 to
~40.

> Because I stumbled upon the Chicken Gazette[1] today, and I'm
> wondering whether it would make sens for Guile to have a similar
> regular source of news?

I think that there doesnt happen enough in Guile land to justify
having an own newspaper. Also such a news and the effort put in
quickly becomes obsolete. Nobody cares for last months news any more.
The effort is better spent elsewhere.

From a casual user point of view, Guile in its current stage of world
domination needs other things first, so if you have time to waste,
consider to sink it into one of these:

* Most importantly, a Wiki.

* A web forum, the ML might be too exotic for "normal users".
  guile-devel can stay a ML, but guile-help should become a
  forum/wiki.

* More programming examples and entry points for newbies.

Considering that guile's (correct me if I'm wrong) goal of being the
"official" high level language of the whole GNU system, and thus to
get more, many more users, it is not that easy to get into.

First, it is a Lisp, which is nowadays exotic enough to scare away a
substantial number of users. Second, it is a Scheme, which is as
exotic in the Lisp universe as Lisp itself is exotic in the global
programming universe. Third, even in the Scheme universe, Guile is a
rather exotic choice compared to Chicken and Racket, so people
managing to get all the way to the Scheme universe are still unlikely
to land in Guile land.

If you dont know a Lisp already, say Elisp or Common Lisp, getting
into Guile by Guile documentation alone is a substantial task, since
the only existing documentation, the manual, is written more in the
terse style of a man page than "Learn Python the hard way" or "Land of
Lisp". It also does not promote Guile as being a general programming
language like Python/Ruby/Perl, but as "just" an extension language
for programs written in more "serious" languages.

* A mighty "batteries included" distribution.

This was adressed also in the Chicken gazette issue you mentioned.
Additionally to getting a online package repository, which is
thankfully being worked on, Guile needs a "batteries included"
distribution similar to the Haskell platform

* http://hackage.haskell.org/platform/

or Lispbox

* http://www.common-lisp.net/project/lispbox/

Which means, it should be available not only for GNU/Linux, but for
all major platforms including the proprietary ones, be an easy,
prepackaged clickety-click install for absolute newbies to be able to
start it, dabble and get comfortable with it, play around with easy
examples, create a GUI (GTK, ncurses) or a web app, etc, in general,
have no entry tresholds. I mean, 2.0 only recently entered sid, and
Ubuntu 11.10 is still shipping 1.8. So the supposed Guile newbie is,
to get started with 2.0, in 2011, still supposed to first build the
whole Guile chain himself to even get started.

If newbies can not start learning with guile, they'll start learning
with python or ruby and then stay there. Not only Guile, but the whole
Lisp ecosystem in general is risking it's future by catering only to
"hackers" and cutting off new blood.

Thanks for reading,
Anton

-- 
NEU: FreePhone - 0ct/min Handyspartarif mit Geld-zurück-Garantie!		
Jetzt informieren: http://www.gmx.net/de/go/freephone

guile user base

[rixed]

Warning, non-technical ramblings incoming.

In your opinion, how does guile user community compare (in size) with other
free schemes user communities? I am under the impression that guile user base
is somewhat smaller than chicken' or racket', but to what extent?  And how
could we roughly figure this out? Also, how well is the user base growing since
v2 came out? Do we have some download stats for instance?

Because I stumbled upon the Chicken Gazette[1] today, and I'm wondering whether
it would make sens for Guile to have a similar regular source of news?

Or maybe I should just draft a first one and see if it draws some interest?

[1]: http://gazette.call-cc.org

Re: guile user base

[Ludovic Courtès]

Hi!

I don’t have any data to answer your question.  It also seems to me that
the Racket and Chicken communities are very active, and have good tools
for that, such as PLaneT and Eggs.

Yet, Guile has been one of the most widely installed Scheme
implementations, which I think is mostly because it started as “the
official GNU extension language”.  This has two implications: a few
widely used applications depend on it (LilyPond, GnuCash, TeXmacs, cgen,
etc.), and socially, it targets free software hackers in general and not
just Schemers or functional programmers (I’m not arguing on whether it’s
successful at that, just that the approach is different from that of
other Schemes.)

FWIW, Debian’s popularity contest shows these ranks (see
<http://popcon.debian.org/by_inst>):

  868   guile-1.8-libs
  15462 racket-common
  24184 libchicken5

Thanks,
Ludo’.

Re: guile user base

[Linas Vepstas]

On 27 November 2011 16:23, Ludovic Courtès <ludo@gnu.org> wrote:
> Hi!
>
> I don’t have any data to answer your question.  It also seems to me that
> the Racket and Chicken communities are very active, and have good tools
> for that, such as PLaneT and Eggs.

The guile projects page is skimpy, because its not maintained: I know
of five projects using guile that are not listed: gnucash, g-wrap,
gnome time-tracker, opencog and guile-db.  gnucash has been using
guile for more than a decade, but is still not listed...

> etc.), and socially, it targets free software hackers in general and not
> just Schemers or functional programmers (I’m not arguing on whether it’s

Socially, it clearly hasn't attracted the kinds of people who insist
on volunteering to run a wiki, a newsletter, etc.  Successful projects
tend to have fan clubs of users who aren't (strong) programmers, but
fall into the class of power users who really want to promote their
favorite hobby. I don't see that guile has any of these.

One of the ways that one gets this sort of attention is brag lists of
things that you think you do really well, especially when the thing is
both arcane and powerful: the sense of leading edge, being created by
really smart people who know their shit, are mind-share leaders
(technically impressive, hip, coool, etc.)   Certainly, the racket
guys have achieved this, with their description of why they changed
the name to 'racket': they've got a bunch of new cool whiz-bang
features that sound impressive.  So, e.g. when I mention scheme in a
conversation, people who more or less don't really know what lisp is
will still say 'oh like racket', because its something they'd heard
of, seen mentioned, and read up on.  Come the day they decide to learn
scheme, for real, they'll pick the one they'd heard of before.

(guile could do this by bragging e.g. that there's a mini-javascript
interpreter or something like that on top of the guile VM and that
would impress people, even though they are neither javascript nor
scheme programmers: but they'd think it was cool, and it proves some
sort of general notion of expressiveness or power.)

> FWIW, Debian’s popularity contest shows these ranks (see
> <http://popcon.debian.org/by_inst>):
>
>  868   guile-1.8-libs
>  15462 racket-common
>  24184 libchicken5

I was going to complain about these numbers before I noted that they
were the *rank* not the number of users. So lower number == more
users.

--linas

Re: guile user base

[Ludovic Courtès]

Hi Linas,

Linas Vepstas <linasvepstas@gmail.com> skribis:

> The guile projects page is skimpy, because its not maintained: I know
> of five projects using guile that are not listed: gnucash, g-wrap,
> gnome time-tracker, opencog and guile-db.  gnucash has been using
> guile for more than a decade, but is still not listed...

Right.  Please don’t hesitate to send new entries (in HTML), preferably
for software that works with 2.0 (though 1.8 is probably OK too), and
I’ll be happy to add them!

Thanks,
Ludo’.

Re: guile user base

[Peter Brett]

ludo@gnu.org (Ludovic Courtès) writes:

> Right.  Please don’t hesitate to send new entries (in HTML), preferably
> for software that works with 2.0 (though 1.8 is probably OK too), and
> I’ll be happy to add them!

I'll start!

<DT><B><A HREF="http://gpleda.org/">gEDA (GPL Electronic Design
Automation)</A></B>
<DD><TABLE>
<TR><TD><I>description</I></TD>
<TD>The gEDA project provids a full GPL-licensed suite of electronic
design tools, including schematic capture, simulation, prototyping and
production.  Guile is used for extensibility and customisation of
many gEDA applications.</TD></TR>
<TR><TD><I>license</I></TD>
<TD>GPL 2 or later.</TD></TR>
</TABLE><P>

gEDA is compatible with Guile 2.x.  Also, this page seems to be written
in HTML 2!?!?

                       Peter

-- 
Peter Brett <peter@peter-b.co.uk>
Remote Sensing Research Group
Surrey Space Centre

Re: guile user base

[Ludovic Courtès]

Peter Brett <peter@peter-b.co.uk> skribis:

> I'll start!

Added, thanks!

Ludo'.

Re: guile user base

[rixed]

Thank you all for your responses.

The target audience of the newsletter I was contemplating being free
software hackers rather than inexperienced programmers, I don't think
a wiki, a stackoverflow-style website nor a web forum would suit ; In
essence because the information that would be provided would be
obsolete quite quickly.

There are certainly many cool things that could be done to attract
newcomers - first and foremost, update the web site as mentioned - but
that's not my intent for many reasons (mostly, I think guile is more
useful as a free software extension tool than as a general programming
learning tool, but this is of topic). What _I_ feel the need of, being
part of the herd of free software programmers inexperienced with scheme,
is a succinct and up to date source of news related to the project, so
that I can:

- see how well or how bad things are going
- what's coming
- what's requiring help
- what other projects use guile and how
- how to make the most of this or that uber-feature
- ideas of new things to try
- examples of common problems and solutions
- explanations on how things work from the inside
- etc...

In short, anything to keep me up to date, give my ideas and knowledge.
I believe this can be useful to people already using guile as well as
attract the attention of potential new users.

A monthly newsletter of the size of the Chicken Gazette looks appropriate
to me. As for the free time I'm willing to spend on this, well, I'm
offering to compile a first edition only in the hope that other people
would volunteer for the next, as :

- I'm not knowledgeable in Guile not Scheme
- I have not many time per week for anything that's not coding (pity my
  children)
- I'm not comfortable with English
- Scheme comes only third in my list of favorite programming languages
  (C and OCaml being the firsts, for those interested)

Anyway, I'm going to try to compile some first newsletter with abstracts
from the guile MLs and IRC channel. I may also bug some people
personally for some more infos. Then I will host the thing, buz about
it a little and see what I catch in my net :-)
Enough talking, we will see then.

Re: guile user base

[Linas Vepstas]

On 28 November 2011 03:42,  <rixed@happyleptic.org> wrote:
>
> - but
> that's not my intent for many reasons (mostly, I think guile is more
> useful as a free software extension tool than as a general programming
> learning tool, but this is of topic).

Well, not really. People get involved in stages, steps, ladders. For
example, someone might want to modify some part of gnucash in some
way.  Maybe a few lines of scheme will do. This step taken, the doors
now open to a deeper understanding of guile. Each of these doors in
turn lead to others.  At some point, the casual user becomes an active
member of the community.

So: I think gnucash represents a non-trivial example of guile as an
extension language. I think it would be excellent if there were
tutorials for how to actually extend gnucash, using guile. As far as I
know, none exist.  Ditto lillypond, geda, or aisleriot, if accounting
is not your bag.

I mean: what's the point of an extension language if no one knows how
to use it to extend anything? This is where the code should be laid
bare: its hard to screw with it when its cloaked by the app.  What are
extension best practices?  What paradigm/api/style can can a guile
programmer expect all guile-using apps to provide?  Some standardized,
easy, obvious way to extend an app, in a fashion that works more or
less the same way for all apps?  All this info is missing.

--linas

Re: guile user base

[Ludovic Courtès]

Hi Cédric,

rixed@happyleptic.org skribis:

> Anyway, I'm going to try to compile some first newsletter with abstracts
> from the guile MLs and IRC channel. I may also bug some people
> personally for some more infos. Then I will host the thing, buz about
> it a little and see what I catch in my net :-)

Sounds cool!

There’s also the possibility of doing something more informal than the
Gazette, such as the “Quarters of the Hurd” (see
<http://hurd.gnu.org/>).  These could be submitted on the Savannah news
system, which gets aggregated to planet.gnu.org.

Speaking of which, it should get aggregated elsewhere, most notably
Planet Scheme.  Anyone knows how to add it there?

Thanks,
Ludo’.