RE: Open Source

["Jonathan Shore" <jshore@example.com>]

Summary:

Technically you are right about the publishing open-source time issue.
Practically and because of the standard I wish to maintain with released
software I do have a time issue.  Longer term I plan to make the useful bits
available as open-source (useful by my definition).

Also, the open-source debate as it applies to my short-term issues is purely
a political and community one.  In the short-term, making our software
open-source is not going to help the company and is unlikely to be useful to
anyone due to the context I will outline later.

Once time permits and as a more general picture develops from our efforts we
will provide as many bits as we think are useful as open-source - with the
standard that we wish to maintain.  I think dumping our sources out  into
the public without having done the due dilligence (documenting,
generalizing, etc) does very little service.


> From: Stephen J. Turnbull [mailto:turnbull@example.com]
> No, as far as I can tell the effort of publishing is negligible.  For
> heaven's sake, surely a firm with domain name "e-shuppan" can put a
> tarball on a web site with negligible effort!  He's already spent more
> effort on this thread than it would take to slap on a license and
> publish the code.


You are right - publishing as-is is negligable.  If the work requires
significant context to be useable or make sense then you have an additional
issue.

I guess what it's coming down to here is that I am unwilling to provide
something as open source at any given time (in my case the short term)
unless I have the time to provide it with a general build env, basic
documentation, generalizing the work (removing it from my specific context),
etc.



> What he seems to object to is the effort of running a bazaar-style
> project.  He keeps referring to the coordination of dispersed
> developers and so on.  To wit:
>
> Nobody except you is proposing a distributed development approach.[1]
> I'm beginning to wonder if that is FUD, or if you've simply been to too
> many ESR seminars.  You're certainly not reading the posts you reply to.

You are right of course.  I have, however, consciously not separated
open-source from distributed development because ultimately this is where an
open-source project tends to lead.  As for seminars, you'll never catch me
at one ;)

Technically, if I am not concerned about the context issue mentioned before,
providing my work as open source does not require an additional time.  On
the other hand if I want this to be useful for others it may require
significant work.

If I provide something as open source I want to be able to provide the
following:

* basic documentation (at least)
* usable in typical environments
* mechanisms to coordinate work (requires a bit of RCS/CVS setup, maybe even
a web-site)
* env simplicity

Some of the work I'm doing at the moment would violate the 2nd and 4th
points.


> What Scott and I want to know is what is the downside of the one-line
> edits I proposed, one to your crontab for regular snapshots and one to
> your release script for releases?  I think that is a valid question.[2]

I could release the source without much effort - but would not be very
usable given the degree to which it is context specific (by context read
hardware, software, and purpose).

There may be 5 other people in the world who might be able to use it as is -
if they even understood what they are looking at.  It is really *that*
specialized.  The specialization is such that personal users would not have
the context (hardware & purpose) to find this useful.

>
> So far it seems unanswerable.  The only presentable[3] reason I can
> come up with is "I don't want to publish anything I don't plan to
> support."  I think that is a valid position to take; personal and
> corporate reputations are at stake.

Not from a support point of view as much as it would be hard to make the
work general enough to be both understood and useful if published as is.

JS