Questionable GNU projects
John Kodis
I was looking through the list of tasks that the GNU project would
like to see people work on, (as reported in the web page at
http://www.gnu.org/prep/tasks.html) and while most of the projects
seem most worthy, there are a few tasks that seem a bit odd for one
reason or another.
For example, in the section of the task list dedicated to the highest
priority projects are a program that implements the Kermit file
transfer protocol. This doesn't seem like the sort of thing that
there's a huge demand for, and if there were, genuine Kermit is
available in source code form with very few restrictions placed on its
use.
Another high priority project is a re-write of the Qt widget library.
While I can understand the need for this, doesn't the existing Gtk
library developed to free the GIMP from its Motif encumbrances pretty
well fill this void?
Similarly, wouldn't the existence of Postgres SQL or GNU's own SQL
database largely eliminate the need for a dBase {2,3} clone?
Lastly, the development of a Unix hosted BBS is listed as a desirable
project. This may have been handy before the near universal
availability of easy internet access, but seems like a somewhat dated
project idea in today's environment.
I hold the GNU project in the highest regard, and don't mean this to
be needlessly inflammatory. However, the priorities reflected in the
tasks list seems a bit out of balance. I wondered if anyone had any
ideas as to what might account for this seemingly odd project
selection.
-- John Kodis.
Todd Graham Lewis
On 10 Jul 1997, John Kodis wrote:
> I wondered if anyone had any ideas as to what might account for this
> seemingly odd project selection.
Age. I'm pretty sure that some of that list dates to the 80s.
--
Todd Graham Lewis Manager of Web Engineering MindSpring Enterprises
(800) 719-4664, x2804 Linux! tle...@mindspring.net
Aydin Edguer
I do not now and have never worked for the GNU Project or the FSF so I cannot
give an official answer. If you really want one, I would recommend contacting
the FSF directly rather than posting to the gnu.misc.discuss newsgroup.
In article <slrn5s8oqg...@kodis.jagunet.com> ko...@jagunet.com writes:
>For example, in the section of the task list dedicated to the highest
>priority projects are a program that implements the Kermit file
>transfer protocol. This doesn't seem like the sort of thing that
>there's a huge demand for, and if there were, genuine Kermit is
>available in source code form with very few restrictions placed on its
>use.
[Note - I think that Columbia University is welcome to use the license of
their choice. I think the work done by the Kermit group is excellent and
I can even understand why they picked the license they did, but...]
"[Genuine] Kermit is available in source" but I might disagree with the
"very few restrictions".
#1 - not all versions are available [without a license agreement] in source
format. In an effort to increase revenue, the Win95 and WinNT versions
of Kermit are not available for free or with source.
#2 - The commercial license prohibits source code changes without the consent
of Columbia University. To quote:
Kermit Development and Distribution
Columbia University Academic Information Systems
New York, NY, USA
March 1996
...
2. The supplier must not modify the Kermit software source code
without consent of Columbia University. If changes are needed,
they must be coordinated through Columbia University so they can
be supported and carried forward in new releases. The supplier
can, of course, produce tailored initialization files, command
files, macros, scripts, tip sheets, and similar material that does
not involve changes to the Kermit software source code.
This prevents other free applications from making use of the kermit protocol
code or the x/y/z modem protocol code or the vt100 emulator code [which is
really quite good].
As I noted above. This is their right, just as it is the right of FSF authors
to use the GPL, but this is why the FSF or GNU Project might be interested in
have a new implementation.
>Another high priority project is a re-write of the Qt widget library.
>While I can understand the need for this, doesn't the existing Gtk
>library developed to free the GIMP from its Motif encumbrances pretty
>well fill this void?
No, it doesn't. The idea is not just to provide a quality widget library,
of which there are many, but to provide a library with a compatible interface
so that free software that was written to use the Qt interface can be freed
of the license restrictions of Qt.
>Similarly, wouldn't the existence of Postgres SQL or GNU's own SQL
>database largely eliminate the need for a dBase {2,3} clone?
No. There are a large number of people who have applications written to
make use of dbase2, dbase3, and dbase4. Things like accounting or inventory
applications, data entry front-ends and report generators. By creating
a dbase clone these users of these applications would be freed of the
license restrictions and gain the chance for buying support which is not
just limited to "fixed in next release". The free software world will also
gain a number of new applications that might fill other requests on the list.
>Lastly, the development of a Unix hosted BBS is listed as a desirable
>project. This may have been handy before the near universal
>availability of easy internet access, but seems like a somewhat dated
>project idea in today's environment.
Amazing enough while the Internet is gaining great favor for certain things,
for others, the local BBS or national BBS (compuserve, aol) is still popular
because they allow small groups of people to get together and get to know
one another and share information. It provides people an opportunity for
finding people that share the same geographic location in addition to (or
opposed to) sharing particular interests.
A BBS also provides a way to give access to a system but still try to limit
the actions a user can take. This can be useful all by itself, even in
a world of client-server computing.
>I hold the GNU project in the highest regard, and don't mean this to
>be needlessly inflammatory. However, the priorities reflected in the
>tasks list seems a bit out of balance.
What do feel would be a "balanced" list? What would you rather see in the
place of the items that you targetted?
>I wondered if anyone had any ideas as to what might account for this
>seemingly odd project selection.
Since the items you selected are hardly the only items on the list (which
is rather long), and since the list can be added to (there is even an e-mail
address provided), I don't see anything odd about finding items that are
not a priority for one person but may be useful to the general free software
community or to a few individuals within the community.
John Kodis
Previously, in response to my questioning the value of some of the
items on the FSF project list, Aydin Edguer wrote:
>What do feel would be a "balanced" list? What would you rather see in the
>place of the items that you targetted?
Well, first let me voice an objection to the term "targeted". It's
not that I feel that any of the projects that I questioned are not
worth doing, only that there seem to be plenty of candidate projects
that would be even more valuable, but which aren't mentioned at all,
let alone being listed in the "High priority" category of desirable
FSF projects.
Specifically, things that I'd find more useful than a Kermit clone or
a Unix BBS would include:
= A WYSIWYG word processor. Ideally one that deals with structured
documents, representing the document as SGML with inclusion of
drawings or pixmaps, but with options to render into text, info, man,
postscript, HTML, and other output formats. If this could double as a
web browser, so much the better.
= A project scheduling package that will accept a list of project
sub-tasks with their interdependencies, and that will generate GNATT
charts and all the other standard project progress reports.
= Some type of map rendering or geographic information system type of
application.
= A serial line analyzer package. This would be a little hardware
do-hickey that sits in a serial data connection and muxes the RX and
TX data streams off to two additional serial input ports. A software
package would sit on these two serial lines, capture and timestamp any
data that comes down the wire, and display it in a fashion that makes
reverse-engineering or debugging of serial protocols easier.
>>I wondered if anyone had any ideas as to what might account for this
>>seemingly odd project selection.
>
>Since the items you selected are hardly the only items on the list (which
>is rather long), and since the list can be added to (there is even an e-mail
>address provided), I don't see anything odd about finding items that are
>not a priority for one person but may be useful to the general free software
>community or to a few individuals within the community.
Quite so. However, given the scarcity of talent available to tackle
any of these projects, it seems beneficial to try to focus the efforts
in the directions that would do the most good for the most people.
This is my only criticism of the list as it now exists -- the
prioritization seems skewed toward projects that would have been
highly desirable a few years ago. It would seem like a better
strategy to skew the list toward projects that will likely be highly
desirable a few years in the future.
BTW, my own guess as to why the list has the makeup that it does is
simply that it hasn't been reviewed and revised in a while.
In any case, thanks for your thoughtful comments. While I agree with
nearly all of them, I still feel that some re-prioritization is in
order.
-- John Kodis.
Elliot Lee
On 12 Jul 1997 14:44:19 GMT, Aydin Edguer <edg...@ces.cwru.edu> wrote:
>>Another high priority project is a re-write of the Qt widget library.
>>While I can understand the need for this, doesn't the existing Gtk
>>library developed to free the GIMP from its Motif encumbrances pretty
>>well fill this void?
>
>No, it doesn't. The idea is not just to provide a quality widget library,
>of which there are many, but to provide a library with a compatible interface
>so that free software that was written to use the Qt interface can be freed
>of the license restrictions of Qt.
I have already done a C++ binding of Gtk. I don't have much time to
spend on making it Qt compatible, but I will certainly do what I can,
so if you are interested, please contact me.
Hope this helps,
-- Elliot http://www.redhat.com/
What's nice about GUI is that you see what you manipulate.
What's bad about GUI is that you can only manipulate what you see.
Thomas Bushnell, n/BSG
ko...@jagunet.com (John Kodis) writes:
> Well, first let me voice an objection to the term "targeted". It's
> not that I feel that any of the projects that I questioned are not
> worth doing, only that there seem to be plenty of candidate projects
> that would be even more valuable, but which aren't mentioned at all,
> let alone being listed in the "High priority" category of desirable
> FSF projects.
The point is that your list of additional projects is great! You
should have suggested it to g...@prep.ai.mit.edu long ago...
The task list is a list of things that seem useful. If there are
gaps, then help us fill them.
I've forwarded your excellent additions on to rms.
Thomas
Ketil Z Malde
ko...@jagunet.com (John Kodis) writes:
> = A WYSIWYG word processor. Ideally one that deals with structured
> documents, representing the document as SGML with inclusion of
> drawings or pixmaps, but with options to render into text, info, man,
> postscript, HTML, and other output formats.
Personally, I find Emacs's (I dare not pronounce the prefixed X in this
newsgroup :-) psgml quite adequate for generating SGML/HTML documents.
What would be needed IMHO is a WYSIWYGgish (who invented this word?
Shoot him before he makes another!) SGML-to-xxxx converter, where I
could adjust margins, select fonts and sizes, scale images etc to my
liking, and then generate xxxx output (where xxxx is some language with
non-ambigous copy output - TeX, dvi, troff, PS, acrobat, I don't really
give a damn, as long as it prints nicely.
Keep the electronic version structured, and free from formatting and
layout garbage.
Of course, this will never win over a hard core MS Word user, but if
people want Word, let them use Word, is my opinion. I could use
something better.
~kzm
--
If I haven't seen further, it is by standing in the footprints of giants
Wolfram Gloger
kdp...@hpmail.lrz-muenchen.de writes:
> It would be enough to write a GPLed replacement for xforms (I hear
> there already is one,but I forgot the name)
wxWindows is an excellent replacement for xforms; it uses C++, but not
any of the newer/more esoteric features, and hence compiles on
practically any C++ compiler (including gcc, even version 2.5). I've
wanted to port LyX to wxWin for a long time, but have lacked the
opportunity...
> IMHO it's
> better for the free software community to contribute to already
> existing free projects to make them usable than to start new ones.
This is _so_ true. wxWindows has been around for more than 4 years,
still people start from scratch again and again to develop GUI
toolkits -- or worse, they base free software on non-free toolkits
(like xforms or Qt), when there are excellent, free alternatives
readily available.
Regards,
Wolfram.
kdp...@hpmail.lrz-muenchen.de
Wolfram Gloger <wm...@dent.med.uni-muenchen.de> writes:
>
> kdp...@hpmail.lrz-muenchen.de writes:
>
> > It would be enough to write a GPLed replacement for xforms (I hear
> > there already is one,but I forgot the name)
>
> wxWindows is an excellent replacement for xforms; it uses C++, but not
> any of the newer/more esoteric features, and hence compiles on
> practically any C++ compiler (including gcc, even version 2.5). I've
> wanted to port LyX to wxWin for a long time, but have lacked the
> opportunity...
I meant a API compatible replacement.
-Andi
Bill Gribble
Wolfram Gloger <wm...@dent.med.uni-muenchen.de> writes:
> This is _so_ true. wxWindows has been around for more than 4 years,
> still people start from scratch again and again to develop GUI
> toolkits -- or worse, they base free software on non-free toolkits
> (like xforms or Qt), when there are excellent, free alternatives
> readily available.
There's a certain virtue to re-inventing wheels if the original
version is designed so that it won't roll :)
I have yet to see the Widget Set That Makes Me Love To Program X.
wxWindows may be nice but it ain't all that. I don't mind people
making new widget sets.
Bill Gribble
Keith Wright
> What would be needed IMHO is a WYSIWYGgish (who invented this word?
> Shoot him before he makes another!)
I believe it was an alleged comedian named Flip Wilson, who had his fifteen
minutes of fame in the late 1960's. He had a television program during which
he would dress in woman's clothes, call him self Geraldine, and wiggle around
squealing "What you see is what you get" in a high pitched voice. Even the
TV-idiots found it got old quickly, and it was canceled after a short time.
The guy who mad the phrase last three decades by applying it to a word
processor certainly has his own circle in hell, but I don't know who
he is.
--
--Keith
This mail message sent by GNU emacs and Linux.
Power to the people. Linux is here.
Food, Shelter, Source code.
Wolfram Gloger
Bill Gribble <gr...@cs.utexas.edu> writes:
wxWindows is not a widget set (since you mention `X' explicitly I
assume you are talking about X `widgets'). In fact, it is available
for platforms (MS Windows, Mac) where there are really no widgets
conceptually.
So I see two complaints that you maybe wanted to raise:
(a) You don't like the wxWin API -- that is fine, but maybe you could
spell out a concrete objection, so that a fix can be considered.
Developments for wxWin 2.0 are underway right now, so it is a good
time to collect suggestions.
(b) You don't like the Look&Feel of the widgets that wxWindows uses.
That means currently (on the X11 platforms), that you dislike
XView _and_ OSF/Motif _and_ and the FWF widgets (a free, efficient,
light-weight replacement for Motif that looks practically the
same), because wxWindows is available for all these backends with
the same general API.
In that latter case, I understand that you don't mind a new widget
set. Personally, I am quite happy with the choices that are already
available. The wheel certainly does roll :).
Regards,
Wolfram.
Vladimir Alexiev
> = A WYSIWYG word processor.
Try XEmacs. Not quite wysiwyg, but workable.
> Ideally one that deals with structured documents, representing the document
> as SGML
The latest psgml mode is great (or is only its html application great?)
> with inclusion of drawings or pixmaps
The hm--html-minor-mode from hm--html-menus has drag and drop insertion of
images and links. XEmacs can display images inline. Not quite wysiwyg, but who
wants that for html anyway? And there's x-symbol for wysiwyg input and editing
of latex math (symbols, sub/superscripts). It (or another similar mode) can
hide the markup, or make it almost invisible, so that it doesn't get in the
way.
There's also enriched-text-mode.
> If this could double as a web browser, so much the better.
Emacs-w3.
>