TODOs.

RoryO roryo at bigpond.net.au
Sat Jun 16 05:03:35 PDT 2001


On Saturday 16 June 2001 20:34, you wrote:
> Yo,
>
> These are some of the (major) issues we need to sort out... i will add
> to this list as i collect them, i suggest you guys do too.
>
>     o Design Documentation
>
>         I'll try and put something together in this regard tonight, but
>         at this point, there isn't alot to say, not without more input
>         from you guys, unless you don't mind me going with whatever i
>         like about the below items in this list;
>
>         You guygs suggested this, and i have no idea where to start,
>         what information should we include? does anyone have any
>         examples? just point me in the right direction and i'll gladly
>         start this task.

Yep, i got me a whole buch of template to sort through.  We should probably 
just start with a basic text one b4 we move to xml or whatever... I'm off to 
bed in a bit, but when i awaken I'll post a basic one - if you haven't beaten 
me to it.  We then start hacking at that for the next couple of days and 
*w00t* the hackers can get their act together.  ;)

>     o Website redesign/information updates.
>
>         This mostly has todo with updating the information on the
>         website, as in, whom is working on what, the Introduction/White
>         paper/Road map, etc... (this will be done most likelly by myself
>         or Gerard, so no worries in this regards)

I'm glad, i know bugger all about html.

>     o Profile Syntax clean up
>
>         This seems to be a topic you guys want to talk about, the
>         profiles aren't perfect right now, they work, but the syntax is
>         terrible and needs alot of work.

You got that in one.  It's more a problem of maintinance than anything else 
at the moment.  If you remove params and such it can be a pain to re-write 
half the profile *exegeration i know* .  We will also need to think about and 
plan for the eventual package management functionality.  Yes HIghoS, I know 
you are not really concerned with it at the moment having MANY more higher 
priorities, but it MUST be considered at the begining or else we are going to 
have to re-write when it comes time to impliment it.

Probably the most anoying thing about it the sytax is the lack of IF... I 
know it could not be implimented in the profiles, but a function of the 
backend could be to check if the file/directory exists B4 trying to act on it 
and them bomb out...  Some though is required here, but you get the drift..

>         I would actually like to see how well a profile would work using
>         more tags and less attributes. (as it has been suggested)

Yeah, i'm going to break the gnome profile down into parts; for those that 
are interested HIghoS is going to host it on quasar for my until we get a CVS 
or some-ting for the profilers... It's just missing a couple of elements... 
be up tomorrow *hopefully*

>
>     o Profile Syntax Documentation
>
>         This should be some of the first documentation written after we
>         have done most of the syntax clean up, it should also be a guide
>         to anyone willing to learn howto create, edit and learn howto
>         write profiles.
>
>     o Profile DTD
>
>         This goes with the syntax clean up, will be needed to be
>         written.
>
>     o IPC/Protocol
>
>         I've trying to start a conversation about what to use, mentioned
>         XML-RPC, no one has bothered to even shoot down the idea, so
>         either you guys don't care or are on vacation.
>         (remenber thou, we want to try and keep this as un-language
>          specific as possible)

Ok, i've done a couple of hours research into the issue and after some 
thought it would seem the XML-PRC.com is the most widely supported/used and 
mature implimentation, so it's got my vote. Good call HIghoS.  ;)

Another issue to be considered is a secure method of communiacting between 
the fe/be.  Authentication of clients and verification of system calls will 
have to be adressed soonest, before coding commences.  From past experiance 
trying to tack security on as an after-thought is never a good idea.

>
>
> I'm going to warn you all right now thou, that if i don't get some
> feedback and input about things we need todo, i will assume that that
> means you guys either don't care or will use whatever i see fit.

Feedback.  We care ;)

> I can't speak for anyone else involed (Gerard, Jason, etc) but we need
> to start somewhere and if you guys keep staying quiet about some of the
> above issues there is nothing i can do except continue on using my best
> judgement.

Your judgement hasn't been too bad as yet.  ;)

> I will, as always, give you guys some time to think and post your
> thoughs and keep answering your questions, but the line has to be drawn
> somewhere, everyone keeps making suggestions and no real work is getting
> done, now.. that isn't a bad thing, as we don't want to jump head first
> into something without knowing what we are getting into, but if we don't
> start writting the design doc soon... well we end up going in circles
> and not getting anything done.
>
> If you guys don't want to use XML and would prefer another method, fine,
> there are a number of different projects doing this a different way and
> you are glad to join them; if you guys want more reasons on why XML
> would be a good idea, just ask.

There are other IPC RPC's out there.... but I would rather use XML than most 
of the other choices.

> I don't mean to sound like an ass, and as always, this is just my own
> thoughs on the matter, but most of the last ~20 posts could have been
> allready answered if we had started doing half of these tasks, but then
> again, perhaps not... I don't mind doing all the work (if it comes down
> to that :) if you guys will help me out and point me in the right
> direction such as giving me some details that should be include in say..
> the Design docs.. i'll get hacking on it, ASAP.
>
> C'mon guys! there was such fire in the language thread, lets put some of
> that to use again, Eh!
-- 
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