[theforum] weo signoff

David Kaufman david at gigawatt.com
Thu Mar 3 20:24:33 CST 2005


Morgan Kelsey <morgan at morgankelsey.com> wrote:

> martin said:
>> 2) the current teo design, while not ideal as an end state, is Good
>> Enough for launch, and we'll look at it again - probably with a
>> competition - after go-live
>
> seeing the time this implementation has taken (i'm not bitching, i
> wish i had time to help too) i don't feel optimistic that this will
> get done before xmas. i'm not comfortable with "good enough for
> launch", we're short-changing ourselves.

I didn't have time to pitch in either, but I have to disagree with you
there.  This project, from inception to implementation consumed only
about 4 four months, those being November though February with, as
usual, the bulk of the progress piled up on the tail end :-)

That's not friggin bad.  It is a month late but I seriously (and
vocally) doubted that it could be done in anything like that
timeframe.  I argued that it was too ambitious to try to rebuild
the front end and the backend at the same time, and change O/S and
change hosting providers.  But they did pull it off, and while it
is not as polished as the current WEO, the polish there is the
culmination of *years* of effort, not months.

And IMO we who did not contribute to this effort should be 
congratulating not nitpicking.  I have to eat my hat, with a side of 
crow, and my own words for dessert.  The Drupal framework and the team 
of overworked volunteers who built this thing have all outperformed my 
expectations.

By delaying the relaunch we're telling these guys they did not do it 
good enough and IMO they did.  They were not trying to implement the 
ultimate redesign.  They implemented an *improved* design on a radically 
different platform and managed to preserve 98% of the structure, content 
and functionality of WEO.

WEO is Dead.  Long live the newborn WEO.  Let's stop hand-wringing over 
what "people might say" and start enjoying what we've achieved, and 
improving what we now have:

 - a remotely admin-able box that evolt rents and no one can take away
 - a truly collaborative platform that we can all have access to
 - a server-class operating system, DB and webapp-framework that we own
 - a best-practices open source configuration that we can all
   afford to duplicate our own dev environments for, recommend to
   clients with a straight face, make affordable fail-over sites for

Even the gearhead techie nerds among us agree that more work is needed 
on the front end.  So lets go live and let the world watch us, as we eat 
our own new dogfood.


>> Therefore, can we please sign this thing off by midnight GMT on
>> Friday. I'm going to assume silence means that you're OK with it, so
>> if you have a burning showstopping reason why not, speak up!
>
> well, those are my reasons for holding off a little....i've never
> relaunched a site for a client that wasn't as good as the original.
> we're not our own worst client, are we?

But evolt is not the client of those doing this development.  If it was,
it would pay it's developers for their hard work.  Evolt is our charity
case, and as such, I think evolt is enjoying a hell of a good deal.

> can we have a vote?

I vote +1 -- Let's launch this thing, warts and all ...Damn the
Torpedoes!


> can finance tell us how long we can afford to pay jeff?

Sure we can but, it's not a question of how long we can afford to pay
Jeff.  Jeff has donated years of hosting to evolt.

That said,

<financial report>
  we still have over two thousand dollars in the bank, so the $50
  Jeff is charging per month could, in the simplest analysis,
  last us a few more *years*.
</financial report>

On Tuesday, February 22, I <david at gigawatt.com> wrote:
>
> I should point out that our monthly Google ad revenue is consistently
> covering the hosting costs of only *one* of our two servers, the
> current production WEO (hosted by Jeff) *or* the new development
> machine (hosted at ServerMatrix).  As long as we maintain both we are
> eroding our bank balance.
>
> So, whatever we do, consider this a vote for doing the thing that gets
> us off the old box sooner than later.

I'm sorry if that recent post of mine was misunderstood to imply that we
were running out of money -- we're not.  We're just going in the wrong
direction :-)

We're "eating into savings" instead of "living on what we earn" because
the Google ads are earning about $80-90/month and our TWO web servers
are costing us $109, combined.  Donations and Cafepress earn us a few
random $10-50 payments per year, so they don't really figure in.

The issue is that we're a nonprofit organization, which means that we
must have a mission, and our revenues must be expended in the
furtherance of that mission, *not* benefit us financially as
individuals.  For instance:

#1 Legally, our revenue cannot financially benefit those who started the
organization, one of whom is Jeff.  While we could argue that he is both
a member and a vendor, and the money paid to him for hosting are evolt's
expenses, not Jeff profiting from evolt, it is advisable for us to avoid
even the appearance of impropriety and use an outside vendor for hosting
services who is in no way associated with evolt.

#2 We must use our financial assets (currently, the result of direct
donations, BEO Google advertising revenue and cafepress merchandise
sales) only to pay our operating expenses and further our mission.
Failure to do results in our having to pay US federal income taxes on
all of the revenue and/or fines penalties and other similar IRS horrors.

But perhaps more importantly...

Jeff has borne an undue share of the evolt's burden for a long time.
Financially, he pays our colocation and bandwidth (the $50 we have
recently begun to pay him only offsets a portion of that);  we rely 100%
on his personal property: 1. the server hardware itself, his Windows 2K
Server O/S, Cold Fusion and MS SQL Server licenses; and his time which
is the most expensive drain, since he is the only person who is able to
perform many of the necessary day-to-day operations tasks (uploading
article images, iirc?), the only person able to put out sysadmin fires,
and one (of only two the members?) with physical access to the machine
when it needs hardware maintenance, a CD in the drive, backups and so
on.

Assuming he never gets hit by a bus, and continues to pay our bandwidth
bills and always uploads our images promptly, and replaces the hard
dives that fail on our (I mean his) webserver, the single-point-of
failure scenario is unacceptable to most of us.  Trusting busses to miss
people and people's servers to always remain free and in good repair,
and people to remain willing and able to donate their time, their money
and their resources to evolt forever has bitten us twice in our
collective ass.

So in as near a unanimous vote as I've ever seen within the evolt
leadership committees, a decision was reached for Evolt to reinvent
our platform, to rent our own home, to redeploy our site, our lists,
our archives -- all of it.

TO REBUILD BACK END.  That was our goal, and it is met.  We have:

1. a new server hosted by a commercial provider with a 24-hour NOC staff
(so we don't need to "feel guilty" asking for a reboot at 2am on a
Tuesday morning) or being angry when the network has problems

2. ported to Linux, so that any of a number of our members, volunteer
SysAdmins, could *divide* the duties and *share* the responsibilities
of remote administration, something that was not possible on WEO's
current situation and configuration -- Jeff had to do everything

3. ported to open-source software for the web server (migrated from IIS5
to Apache), Application Framework (Cold Fusion -> Drupal), the database
server (MS SQL Server -> MySQL), and of course the O/S (Win 2k ->
Linux), saving money (only prudent for a nonprofit organization),
increasing reliability, and being able to host evolt.org on our own
dime legally.

These were the goals of this redevelopment effort, and it has succeeded
in achieving each of them.

I think redeploying on a commercial service provider, saving the
thousands of dollars in the licensing fees we'd have to pay to host a
Windows box (configured like the current WEO), freeing ourselves us
from the dual legal tangles of paying one of our own members to host
us, and "sharing" his expensive software licenses, is a Big Deal and
the fact that the site works with just a few rough edges means that a
relaunch of evolt this week would be a shining success story.

No one has claimed it is perfect.  If anyone thinks that the rough edges
are show-stoppers, they should at the very least post detailed bug
reports about them.  The team has been very responsive about fixing
known issues that are reported clearly and reproducibly.

Did I mention I vote for a +1 on the signoff?

I say let 'er rip!

-dave




More information about the theforum mailing list