[Rpm-metadata] status questions

[Malcolm Tredinnick]

On Tue, 2003-11-04 at 17:31, seth vidal wrote:
> Hi everyone,
>  I was just wondering what people think of what we've come up with so
> far. Does it seem workable for all the various projects/programs? Are
> you comfortable with the file layout? 

I like it, but I have no corporate or commercial involvement, so I can
afford to take the academic's view. :-)

> here are the things I've left on my todo list:
>  - finalize dtds for filelists and 'other' file

I meant to mention yesterday that I can help you with the DTDs if you
need it. I understand what Daniel is talking about with the namespace
attributes. (But see below for my timing -- it won't happen until the
weekend.)

>  - determine shape/layout of the releases/channels/whatever you want to
> call it file that informs about the rest of this data
>  - versioning for this format so we can tag things as 1.0/2.0/etc in the
> future.
>  - explanation of decision documentation (malcolm, what things do you
> need?)

Another ten hours a day. :-)

Seriously, this week and last week are business as usual (i.e. very
busy) at work, so I have been going home and crashing, rather than
working on "stuff". I am taking a week off next week to catch up on some
of my open source stuff that is piling up and this documentation is on
the list. I will try and do it early in the week.

I think I am fine as far as material to extract the information from. I
have sorted through the archives and extracted out the non-obsoleted
decisions. I can read your code and Jeff Licquia's (my day job involves
Python, so *really* no problems there). Once I have written all this
down, the fun will start because people here will be able to find places
where I have misunderstood what is going on. But let's fall off that
bridge when we come to it.

[...]
> I have this delusion about maybe getting a release of this format out
> for public critique and consumption by the end of november, but I'd
> rather not rush anything if it isn't ready.
> 
> Also if this works out it might be useful to propose that this format be
> given some sort of lsb-status (eventually) - I think Jeff J. originally
> suggested this.

That should be .. um ... "interesting", given the LSB belief in
dependency specification. Are they likely to bless a format that
specifically accomodates for non-LSB compliant packages? But it would be
worth it if we could do it, I think.

Cheers,
Malcolm