syntax, ALFS, nALFS -- versioning

CoreyCox CoreyCox at linuxfreemail.com
Fri Aug 10 10:26:56 PDT 2001


Neven Has wrote:

>
>I'm not sure why you think that's needed. Like I wrote in my reply to Jesse,
>the only standards that we should create and follow are those needed for
>programs and profiles to work with each other.
>I don't think that versioning of various different software and profiles 
>that are now present, or will be in the future, should be enforced.
>
>I look at all this (both licenses issue and versioning) this way:
>
>Think about IRC for example. There is a strictly defined protocol that
>surrounds it. That protocol is used by a bunch of different clients and
>servers which (because of the protocol) communicate very well.
>
>But those clients and servers don't have to have the same license, nor
>use the same versioning system.
>There could be 1.0c18, BSD, 1.2.3, GPL, 990212, or even 010808-fix1. ;)
>
>IMHO, we should limit ourselves to just setting the standards that we really
>need - nothing more, nothing less.
>
>
>Neven
>
Alright, I agree with you on most of this.  My main point is about the 
profile versioning actually.  Currently we are working with "1.0" as the 
version number written into the alfs profiles.  Except your profile 
which has the version listed as what LFS system it is based on.  I just 
think we need ot decide what this attribute is going to be used for and 
what numbering system we are going to use for the syntax.  Here is what 
I think:  First I think the version= attribute should refer to the 
version of the syntax that the profile is written in.  This way the ALFS 
agent could check that and know whether it can interpret it or not - or 
if we get really sophisticated in the future it could allow backwards 
compatibilty.  Second I think we need to decide on a versioning system 
for the syntax, and start using it.  That way we dont' back ourselves 
into a corner when no one can figure out which profiles are suposed to 
wokr on which versions of interpreter.

The next thing I'd like to see is a little different.  This I think 
would be very useful for both developement of profiles and within the 
interpreters themselves.  This would be a module to test for validity 
against a DTD.  I know we dont' currently have one - but I could knock 
one out real quick with the new syntax real quick - or anyone else could 
do it too.  Then this module could be used to test profiles before the 
build process - further it would test whether we are really using the 
XMLcorrectly.  Let me know what you think.

Byt he way I do have those profiles done now if anyone wants them.  I'll 
post to the list tonight - unless someones in a real hurry!

--
Corey Cox.

-- 
Unsubscribe: send email to alfs-discuss-request at linuxfromscratch.org
and put unsubscribe in the subject header of the message



More information about the alfs-discuss mailing list