[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: UNIX 98 (Was: Re: Standard libc...)

On Thu, 20 Aug 1998, Evan Leibovitch wrote:

> On Wed, 19 Aug 1998, Jakob 'sparky' Kaivo wrote:
> > I hope noone minds, but seeing as I've already projected myself into the
> > position, I'm going to declare myself the 'Guy Making Sure That LSB/LCS Is
> > UNIX 98 Compliant (GMSTLIUC).'
> Gee, I'd thought that was Ian Nandhra's job :-P.

I must have missed that announcement. I'll step down if he wants me to.
But, as he hasn't contacted me yet, it doesn't seem that he wants the job
back. ;)

> You may have more of a task ahead than you think. There are some elements
> of the Single Unix Stsndard -- such as STREAMS -- that will likely be hard
> to get pushed into a standard Linux environment (just ask Caldera :-).

Oy, I forgot about that mess!

> Likewise, there are some significant parts of the LSB -- the init-script
> system, for instance -- which are considered part of CPU-specific ABI
> specs and are thus outside the scope of the SUS. 

Yes, they are outside of the scope of the SUS. That's perfectly fine. I
just think it is important for the things that *do* fall within the scope
of the SUS are compliant (or can be easily made so, or are documented as
to what the difference is and why).

> The best benefit of a "TOG watchdog" role is, to make sure there are no
> gratuitous variations that can be easily fixed, and to document those
> differences between the SUS and Linux where they can't be easily
> reconciled.
> At the LI meeting where the LSB was formed, the issue of standards came
> up; the consensus I recall was that strict conformance may not be a goal,
> but at very least a detailed exception list is necessary.
> (Ie, "the LSB conforms to TOG standards document XYZ except for <list of
> specific differences>") 

This is a very good suggestion.

> It was agreed (someone corrent me if I got the
> wrong impression) that conformance was desirable but not critical; that
> a well-documented list of exceptions between the LSB and SUS would
> suffice. I'll gladly offer to help compile that exception list.

Thanks. Well, I'll be ordering the bound set of the specification soon,
and then will start going over things with a fine toothed comb. In the
meantime, I think everyone should continue as normal, and if I notice any
glaring omissions and/or contradictions, I will try to point them out.

| Jakob 'sparky' Kaivo        |          jake@nodomainname.net |
| NoDomainName Networks       |    http://www.nodomainname.net |

Reply to: