[RFC/P] Port System Re-Engineering

[Aryeh M. Friedman]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Chuck Robey wrote:
> Paul Schmehl wrote:
>> Here's a hint that would help a *ton* of users.  Don't try to
>> install a port until your ports tree is up to date.  Completely
>> up to date - as is, run portsnap or cvs or cvsup *first*, *then*
>> try to install your port.
>>>
>>> I have several possible solutions (contact me privately if you
>>> want more detail) but am purposely not stating them publically
>>> so as not to taint the survey any more then it needs to be.
>>>
>> This is the part I don't get.  If you have suggestions, post
>> them. Post the code that implements your suggestions.  *Then*
>> people can evaluate whether or not your suggestions add value to
>> the ports system.
>>
>> Why the silly games?  As I read them, this seems to be the
>> primary objection of all the people responding who have
>> @freebsd.org in their email address.  They've heard it all
>> before, but they know that actions speak much louder than words.
>> If you say "the implementation of foo is flawed", and then you
>> post code that, IYO, improves it, people with experience and
>> knowledge can review it and say, "Hey, nice idea" or "sorry, your
>> code would break ports and here's why".
>>
>> Without the code, all the surveys and gesticulations in this
>> tread accomplish little except to irritate people.
>>
>
> Why the silly games?  I get the feeling that Aryeh is honestly not
> understanding that he's trying to change the basic way that things
> get done in FreeBSD.  He doesn't see that.  In industry, first a
> decision is made that a market exists for such and such, then a
> study is made as to what could be done realistically.  We don't
> operate that way.

If finding the market is all they do but they don't follow the whole
process to produce code then they completely missed the point... the
idea of "modern" (almost anything that post-dates v7 of unix) looks at
how to do the whole process systematically... thats the idea here....
doing a market survey then doing no code is as worse as making random
tweaks for no apparent reason (i.e. you have no idea if they are needed)
>
> What we're all afraid of, Aryeh, is that you're going to run off
> with your poll of what you believe is needed (when we haven't even
> agreed that anything is needed) and you'll code something up, under
>  the completely wrong misapprehension that if you code something up
>  that does what the poll results said, it would get added in, pal,
> that's totally, totally false, you can ask any committer whatever,
> you will never get any apriori agreement on the adding of code, no
> matter what, until we can see the code.  This has been endlessly
> argued in the past, and folks have certainly left FreeBSD over it,
> but it will not change.
This is why I a asked informally for a p4 account (the person I asked
should be asking formally on my behalf soon)... first step is to
establish some baseline about what the project encompasses and what it
doesn't and thats the stage we are at right now...
>
> If you can't see that, then we will remain at loggerheads.  If you
> can see that, then quit asking folks to agree on stuff without
> showing us code.  I don't care how much research you do on what is
> needed, you will never change that fact, all you're going to do is
> trigger knee-jerk reactions from folks who have been *very highly*
> sensitized by prior attempts to change that rule.  It's not gonna
> happen, and you strongly seem to be trying an end-run around it.
> If you honestly aren't, then you need to do a better job of
> convincing folks of that fact.

BTW who ever said I am saying there is something wrong with the
current method of doing freebsd development: I am and the complaint is
we are far to much of coders and not enough of sit back and find the
real problem before solving "it".


- --
Aryeh M. Friedman
FloSoft Systems
Developer, not business, friendly
http://www.flosoft-systems.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.4 (FreeBSD)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHVIHl358R5LPuPvsRAvp7AKDCLBwRCT4onjIe0L/THVFuIMUe4wCeIaH7
mxxoLXxtMD/KAGd/eR8zHJU=
=fUWb
-----END PGP SIGNATURE-----