[theforum] movement for redesign

David Kaufman david at gigawatt.com
Mon Aug 11 20:30:22 CDT 2003

David A. Ulevitch <davidu at everydns.net> wrote...
>> we really need ... exposure to the inner guts of weo, even just
>> from a many-eyes perspective.
> Agreed.  This is critical and *must* be reality.

human fault tolerance?  how much is a redundant array of inexpensive Jeffs?

>> If this means sticking with or migrating from the win2k/mssql/cfm
>> platform, so be it, but i'm sure a full scope and design as part
>> of the buildup to the codevolt will help us assist in ensuring we
>> can *all* use the right tools for the job.
> I can provide any open source resources required (linux, apache, perl,
> php, postgresql, mysql, etc) on a fairly dedicated box in a secure
> colocation facility.  This machine has access to backup machines for
> backups as well as cold-failover.

this is at community colo?  wouldn't weo's bandwidth requirements be an
issue?  or was the bulk of the bandwidth expense on mailing lists and the
browser archive?

regardless, it seems sensible to have such a failover site on a different
box, on a different network, so that if the main site's box or network goes
haywire (or if Jeff gets hit by a bus and no one else has the access
privledges + know-how to step in) the backup can become the frontup in a
hurry, by David just punching the big red "press only in emergency" DNS

> I can provide CVS access to developers and anything else required for
> a coordinated development effort.

that does seem to be the professional platform of choice.  the last two
development shops i've worked for have moved from "using" (playing with) cvs
to mandating all development be done in cvs dev environments.  at both,
nothing is ever touched "by hand" on the production sites -- each
developer's changes are checked in onto a nice, sane version-controlled
branch that is then used to update, if necessary resolve merge conflicts and
test on either a staging-site virtual host or the production site (for
setups.  it's a nice feeling to catch a show-stopping bug, typo or broken
on a staging site before it goes live.  is there any sort of staging done
with weo or is development done on the live site?

i'm +100 on the open source route.  "open migrating" can be fun. in the
last year i was fortunate enough to get to migrate a handful of large sites
from ColdFusion to PHP, Netscape to Apache, Oracle to MySQL, Solaris to
Linux, and am beginning an Exchange to Qmail+LDAP.  kinda makes us feel
like masked banditos, stealing revenue from proprietary software publishing
companies, and returning it to our clients :-)

i think, however, that many threads of messages here, denouncing evil (my
own evangelical posts included), do not a contribution make.

not compared to the contribution of time and (correct me if i'm wrong, free
hosting) that Jeff makes.  so if he is most comfortable donating his time
and effort and hosting services in the form of ColdFusion... perhaps any of
us other would-be contributors should set up a development-only box (perhaps
even ColdFusion for Linux, at leatst to start with...) with SSH logins and
CVS dev environments to allow multiple developers acces to a single
site.  then as soon as useful, well-tested, changes begin to evolve, mutate,
merge and propagate into that staging site, easy updates can be shipped off
to weo that Jeff could roll out as easily as typing "cvs up" (or if not,
at least click+unzip).

the interesting thing will not be so much whether or not enough volunteers
*offer* to participate and contribute to the development, but whether we
can even come close enough to matching Jeff's (reportedly Herculean) efforts
to justify the approach.


More information about the theforum mailing list