Submitted for your approval: White paper draft

Evan Leibovitch evan at starnix.com
Wed Apr 5 10:00:59 PDT 2000


On Wed, 5 Apr 2000, Ralf Flaxa wrote:

> You will see the fruits of it soon, so if you want to help then do rather
> play the advocate for it than questioning it publically.

If I want to question it publicly, I have many vehicles to do so that are
more public than this list. I say these things here and *not* publicly in
the interest of helping the LSB process.

> I appreciate your input and comments, but I feel they have an
> "undertone" of pessimism.

No pessimism is intended. I believe strongly in the value of the LSB and
hope it will be widely adopted, as I have from the day the LSB was
created. I understand that it is heavily supported, as it should be, by
every distribution, and I expect that most distributions will ensure that
LSB compliance is highly publicized.

I'm merely stating what I see in my observations of the landscape. Many
developers are bringing apps to Linux without waiting for the LSB, so they
must already target one or more "standard" platforms (sometimes refusing
to support others). The worst example of that was retail boxes on store
shelves I saw that boldly said "CodeWarrior for Red Hat", but I have seen
others too.

It appears at this point that (most) ISVs who are developing apps for
Linux (who don't do strategic relationships with distros) are, by default,
doing versions for Red Hat, maybe one or two others, and hoping it'll work
on the rest. Even in the free software world (in download areas such as
'rpmfind'), when binary packages exist for free software, more often than
not they're Red Hat RPMs

If this trend continues, and has the opportunity to get ingrained as "the
way to do things", then a defacto standard will emerge with which the LSB
will have to compete (especially if Red Hat is not as aggressive at
promoting LSB as the other distros).

This is not stated as a threat or prediction of doom -- just two main
points:

1) Even if LSB did not exist, a standard would emerge for ISVs making
   apps for Linux, but it would emerge by unplanned attrition rather than
   consensus. Such a standard would likely target a front-runner platform
   and force everyone else to maintain compatibility. The absense of LSB
   would not prevent a such defacto standard from occuring anyway, but it
   would take a form that most distribution vendors (everyone except the
   front-runner, actually) will not find appealing.

2) There *is* a time issue at hand. In the absense of LSB, ISVs are already
   being forced to create an ad-hoc "standard" of their own. The longer
   LSB takes to arrive, the harder it will be to overcome the inertia of
   what the ISVs will have already done.


- Evan




More information about the lsb-discuss mailing list