ALFS Progress (was: Re: bash: ./configure: No such file or directory)

Barry barry at hartford.uconn.edu
Fri Dec 29 11:19:37 PST 2000


OK, time for my opinion...

better late than never... sorry, I was going through a 500 e-mail
archive that was sitting on a PC I haven't gotten to use for a week or
so...things have been VERY hectic at work...I'm writing this on my lunch
break.  Hopefully, this hecticness should be over by the middle of next
week...then I can get back to some more LFS based development...

I think that my thoughts pretty  much mirror those of the other
respondants to this thread...with a few additions.

I, personally, have looked from the perspective of learning LFS for the
time being.  I don't want to get too involved until I've worked out all
the little things that I haven't yet figured out... I feel that I'm
close enough now to start work on it... I know Bash shell scripting,
pascal (turbo, gpc, and freepascal flavors) which will be pretty much
useless here, HTML, Perl (I'm VERY rusty... it's been years since I did
any Perl, but this is a good reason to refresh myself), and C which I
just really started to learn...

I'm willing to learn anything else and XML is not out of the question
for this...but it will take some time....

now, dispensing with the question of my ability to assist in this
endeavour, I will interject some basic observations.  I want to
illuminate the fact that these are not criticisms, but simply facts as I
see them.  They have more to do with the stage that ALFS is in than they
do with any doomsday style critiques that other people may give...

1) Accessibility is necessity.  Once the project reaches a certain
point, then it should be released to the public.  And this release must
be consistant.  If it's not, people simply won't bother.  I know that
there have been many pitfalls in building the ALFS, but to say that the
ALFS profile and any information concerning the ALFS project has been
easily accessible over the past 6 monthes would be an overstatement of
the facts.  Now, it would be very easy for people simply to drop
questions to the mailing list, and anyone (including myself) who has not
done so is so guilty.  But I believe that this is more an indicator of
the maturity of the project more than it's overall scope.  There is no
web site, beyond a page with 4 or so lines of text and 4 links.  3 of
these links point to directories with nothing in them and the 4th is a
link to the profile which, as good as it is, is likely not useful to
everyone involved or does not fit their purposes entirely.  Again, this
has more to do with the fact that there really isn't much work being
done on ALFS right now besides what the three of you are doing.  This
will change in time... I hope to outline how it may be accelerated in
the next few suggestions/observations...

2) Direction is a goal.  If the project has no goal other than to
automate what most people are already doing in their spare time in a
manual fashion because they want to do it manually, then the project
will likely not attract the number of interested parties that you wish
would come forward, and those who are interested will not feel compelled
to work on the project.  The problem, though, is not that the project
has no goals.  I feel that the goals that the project has are venerable,
to say the least.  And many of the usage intentions that people have
surrounding ALFS are even loftier than the initial scope of the
project.  Ideas such as building the perfect LSB compliant system,
etc... Which are necessary for the advancement of Linux - can be founded
on the work that is going on here now.  But developers need a reason to
do something.  White papers, newsletters, a mission statement...  All of
these can be tools which can enliven the public and give energy to a
movement.  These are the provokers which made Linux and the Open Source
community powerful in the eyes of the world.  The code is the important,
tangible resource... but to be successful we must also be statesmen.  To
have a boat is one thing, to go places in it is another, but to expect
people to get onto the boat, you have to have a widely publicized,
pre-determined destination... 

3) Divide and conquer.  The ALFS should be split into coherent parts. 
This will be necessary for it to develop beyond the point that it is
currently.  These parts would, most likely, be more than simply a back
and front end.  In writing my initial shell script for installing LFS
automatically on christmas day, I seperated the LFS into 3 stages.  I
may even seperate it into a fourth.  But for an automated project, I
would suggest a modular approach which would keep bug testing,
development, and further advancement of the project easily managable.  I
would split this up along both a stage axis and an axis for front and
back ends... Let me take a look at this to see how I feel that it could
best be done... I'll write to the list when I feel that I've come up
with a good solution and when I can elaborate more on this...

4) Once the project is divided into coherent parts, the project will be
easier to manage and add to.  Rather than trying to manage people making
additions to a larger product, a smaller modular design will allow for
quicker development on the various pieces.  Keep IPC simple and
standard, that way modules can be modified and replaced without having
to make too many changes to the other modules.  How we do this will
depend largely on how we divide the project.  This will also be
necessary because of the nature of the work.  As we all know, this work
is voluntary, and that means that we have to structure the program in
such a way as to enable quick fixes and an accelerated development
cycle...

5) Don't give up.  I've seen too many promising projects get dumped over
the past 6 years.  I don't intend on seeing this one fall.  You guys are
doing a great job.  We just need to pick up some of the pieces and start
building, that's all :)... It will happen, it simply takes time.  And
Jesse, damnit, you've got leader oozing from you from what I've seen. 
Stop beating yourself up over this :)

OK... phew... hehh... :) 

I hope I haven't pissed anyone off...it wasn't what I meant if I did...
So, I guess this means you guys can count me in.  I'll back up my own
points of view here with some content... hopefully over the weekend. 
There's going to be a major snowstorm here in CT starting tommorrow.  I
can think of no better time to start ... :)  



Jesse Tie Ten Quee wrote:
> 
> Yo,
> 
> Sorry for the double post, but i consider this important enough that
> lfs-apps and alfs-discuss should hear about it/comment/rant/etc
> 
> On Tue, Dec 12, 2000 at 02:44:43PM -0500, Gerard Beekmans wrote:
> > There aren't as many people actively working on ALFS as me and Jesse had
> > orignally hoped for (hopefully Jesse hasn't regretted to take the offer to
> > lead ALFS yet ;) We're still expecting that more people will start helping
> > out once something that works has been released. But it's easy going. I
> > myself have to split my time between LFS and ALFS, but I have made a decision
> > to do no coding on ALFS, or at least to the very minimum. I just know what
> > will happen once I start coding. Nobody else codes so I take over since
> > "somebody gotto do it". I just can't do that. This means that currently Bryan
> > (our employer) is working on a perl backend (but he has to run a company as
> > well so he has not all the time in the world to work on it understandably)
> > and Jesse works on whatever he works on (miscelaneous things like backend,
> > profile, racks his brain on IPC and so forth). What am I doing for ALFS? Not
> > a whole lot to be honest, I'm not the project leader (well I'm the 'global'
> > leader if you will. I oversee it but am not leading the ALFS subproject).
> > I've worked on the profile and am still waiting to improve it (after bryan's
> > done with his modifications to the profile).
> 
> I have always been willing to lead ALFS...but i'm not prepared todo so,
> not untill those that are invovled (Gerard, Bryan, alfs-discuss) agree,
> that i should take on the responsibility of leading the project.
> 
> Why? lets all be honest... because i wonder if i am the right person to
> take on that position, because i wonder if i can do everything that will
> be asked, i wonder if i have enough enough patience to really lead, i
> wonder about my future.
> 
> ALFS has been a big boat going no where for the last 6 months, don't get
> me wrong, there have been two key things (syntax and profile) that are
> almost [1]finished, but not as much as what should or would have been done,
> if there has been someone leading the project.
> 
> I've been thinking about alot of things in the last month.. especially
> as on the '17 of this month, i'll be turning 18, i'm still very young ;)
> 
> > If anybody here is interested in helping out just subscribe to alfs-discuss
> > and let us know what you are able to do (programming wise). We still need
> > quite a bit of help
> 
> It doesn't even have todo with programming, we _really_ need a website
> done, even if that means backporting the LFS website (which i'll
> problably do, if nothing comes up)
> 
> Along with a million other small things, like doing some documentation
> work on the DocBook SGML Syntax, a little PR for the project, getting
> more users and developers interested, helping out in the design/process
> for the IPC.. money, cars, woman and your soul, j/k ;)
> 
> We only have three people working on ALFS right now, Gerard, Bryan and
> Myself, and nothing has been happening on the mailing list, so this is
> what i'm going todo to get a bit more action going under ALFS, untill i
> hear a better way;
> 
> There has been a HUGE number of LFS'ers asking about a way to install
> LFS without having to have an installed Distro first, so i'm going to
> [2]finish the ALFS Bootable subproject.
> 
> At least then, we have a part of the project everyone will find handy to
> have and hopefully we can gain some developers...
> 
> Dunno anymore, i'm sad about ALFS at this point, there has been lots of
> ppl that have said they want to help out, or the project is cool, but so
> far it has been just that, talk.
> 
> So, anyways... i'm going to send this now, before i edit it again and
> rant more ;)
> 
> Please give your two bits of advice, even if you don't know alot about
> ALFS or those invovled....... some feedback is needed, good and bad.
> 
> [1] They are at a production level now, nothing is ever finished, but
> they are usable ;)
> 
> [2] Get at a usable level and release it under ALFS.
> 
> -
> Jesse Tie Ten Quee - highos at highos dot com
> 
> --
> Unsubscribe: send email to alfs-discuss-request at linuxfromscratch.org
> and put unsubscribe in the subject header of the message

-- 
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