Suggestions for dealing with irrtoolset's problems

Jonathan Day imipak at
Mon Nov 7 17:41:58 UTC 2005


A few thoughts with regards the situation with
irrtoolset. As I understand it, ISC doesn't have the
resources to really maintain irrtoolset beyond basic
maintenance, at present, and don't have the manpower
to re-engineer the software using more modern tools
and techniques.

My first suggestion would be to fork irrtoolset, with
both forks under the control of ISC. Have one fork as
the "official" ISC release, which would be maintained
on an as-able basis. The other version would be
available using a source-code management system, such
as Subversion, and would be more of a playground for
people to try patches and experiment. The ISC could
then collect a "best of" set of patches to use in the
official version.

My second suggestion would be to have the Open Source
community as a whole work on producing an initial
modern irrtoolset, which would get round the limited
resources of the ISC. There would be several ways to
do this. The simplest would be to produce a functional
specification, a user interface specification and a
network I/O specification. With those, it would be
relatively easy for a programmer to develop the basic

My final point is that relatively few people know
about irrtoolset. At all. Of those, fewer know that
the ISC has the software and fewer yet know that it is
being maintained at all. The problems with the mailing
list archives don't help. Although patches exist,
there has been no new release since the ISC took over
the software. It's not even listed on the front page
as a project the ISC has.

I fully understand if the lack of a clear mention is
because the software still needs work, but because
resources are limited, I don't see any good option
other than to build interest amongst those who might
be persuaded to help out. Otherwise it's just a
stalemate situation, which isn't good for anyone.

It is not clear how to have both momentum and high
coding standards, but (through no fault of the ISC) at
present there are neither. The problem, then, is how
to acquire both - which is why I'm suggesting that any
"rapid development" version be kept distinct from the
"stable, official" version. Have the momentum in one
version and the high standards in the other.

Other ways to achieve the same results certainly
exist, these aren't the only (and probably not even
the best) ways to improve the situation. They're
intended to encourage the ISC to look at alternative
methods and to at least provide the ISC with
possibilities that would not require additional
resources but which might result in irrtoolset being
more actively - and more visibly - developed and used.

Start your day with Yahoo! - Make it your home page!

More information about the irrtoolset mailing list