(In response to Ian Rogers).
Hi Ian,
Thanks for the comments.
> I enjoy following this group, though mostly out of nostalgia as I have too
> many other projects going on to contribute much right now, but I feel
> compelled to stick my oar in on this issue [and it's turned into a bit of
> an essay, sorry].
>
> I must have missed the decision to change the version number from 15.53 to
> 15.6 (and if anyone can dig up the rationale behind it I'd be very
> grateful) but it seems a bit silly and very broken for an OSS project!
As you have noticed, there has been no consistent numbering scheme since
poplog became available at the Birmingham site. That's because in
contrast to the previous situation where a development and support team
(people at Sussex and ISL) looked after poplog, there is now only one
person here doing it in miniscule amounts of free time, namely me, and
it has all been done in a very ad hoc way, just to get something going,
partly because I was then (1999) totally unaware of numbering/naming
conventions for downloadable packages.
I made a few attempts to get funds to hire someone to help rationalise
and revitalise poplog but failed. So I've continued just fitting it in
and from time to time installing fixes or extensions that other people
give me, and from time to time building 'complete' packages. Originally
there were only for students and collaborators then eventually they
became the default linux distribution, because others wanted the benefit
of the easy installation scripts I had prepared for our students.
I did not wish to have to keep editing files that pointed to dowloadable
systems so while installing minor fixes (which is what most of them have
been) I've kept a fixed file name as a symbolic link to the latest
version of the package, which might have a different name here.
Also the shell scripts for building most of the tar files include in
many cases a command to use 'touch' to add an empty file whose name
specifies the date of construction of the package. It got more complex
as packages started getting nested: e.g. for the benefit of students
here and our collaborators I started producing a package that contained
several tar files, one of them being a core poplog tar file, others
being various extensions from Birmingham, Sussex and elsewhere.
Since there is no process of ongoing development on the strand of poplog
that I look after, just minor extensions, revisions, etc. when it seems
appropriate, there is no schedule of releases. Some of the updates to
the main package came because one of the included packages e.g. rclib,
or popvision, had a small change.
Then around December last year I found some time to start reorganising
and re-packaging the files,
- partly because there was a lot of junk (e.g. out of date teaching
materials and some ancient demo programs that no longer worked) in the
package that had been distributed by ISL,
- partly because other things had been treated as extras for some time
which were being fetched by everyone by default (except perhaps some
experts), and so should become part of the standard bundle
- partly because there were things lumped together in the help, teach,
lib and auto directories that needed to be separated out (e.g. AI
teaching materials should not have been in with the poplog/pop11/ved
teaching materials).
I also wanted to stop using the $poplocal/local directory as part of the
distribution package because I felt that should be left for end users,
especially for places using a shared poplog installation for teaching or
research.
At that point I was talked out of calling the revised version 16 because
the change was merely one of organisation, not anything dee.
Anyhow I regard everything I do as a stop-gap while other people work on
the long term solution which is OpenPoplog, based at sourceforge, not
here in Birmingham.
That project is going to include *major* changes, including
re-designing the default graphical interface tools so that there is
something that works across all platforms (unlike the X window based
version that works only on unix derivatives, e.g. linux, and eventually
OSX, though I believe it was also ported to VMS poplog) and also
reorganising the whole file-system to use far more sophisticated and
principled software engineering tools for generating different
configurations, tracking changes, etc. (Jeff Best is designing all
the management tools I believe.)
I don't know how long it will take before a robust usable system emerges
>from the OpenPoplog effort (to which my only contribution is making sure
that there's a central repository of bug-fixes and enhancements to the
bham linux poplog system so that when the time comes they can be
transferred to OpenPoplog).
Also if people produce ports or version specific documentation, I list
them in
http://www.cs.bham.ac.uk/research/poplog/freepoplog.html
and also make them available at the bham site. That has included
Steve Isard's mini-poplog, various things for debian and freebsd,
and the two ports done by Waldek Hebisch, for HPUX and for
AMD64(Opteron) still in progress.
When the OSX port is done I would expect to make a copy available here
too.
I don't have time to keep the different versions totally consistent.
Eg solaris poplog has lagged behind. (I am not even sure I can rebuild
it here now.)
A better organised person would use CVS or some such thing and
maintain a history of diff files, etc.
As Roger commented, I found the recent reorganisation required far more
effort than I had anticipated so perhaps when that has settled down and
file headers have been corrected to fit new locations, (which will
depend on my writing pop11 scripts to do it) etc. I should give it a new
version name, perhaps V16 or V16-0-0 or some such thing. There will be
an option to convert it to AMD64 using Waldek's patch and build scripts
(though that port is not yet fully functional).
So the system you are commenting on is not the usual kind of ongoing
multi-person effort. It's not even a even a one-person effort. It's
perhaps 1/50th of a person effort (except for occasional periods when I
or others spend a lot of time on some problem). And I don't see it
changing often enough, significantly enough for version numbering to be
an important issue.
I hope that all makes sense.
If someone wants to do some of the things I am not doing that's fine
with me, but I suspect everyone is too busy.
Aaron
|