:: 06-Nov-1998 01:39 (Friday) ::

OK, Here’s the deal…

I’ve finished re-running the 4-Nov logfiles and the numbers came in
higher then they did previously, so that much looks really good. Those
of you who were missing blocks should now have a more appropriate count
for 4-Nov.

However, there is still the issue of the doubled 3-Nov data. This
seems to just happen from time to time, generally around the beginning
of a month, and I’m at a loss as to how it happens. Fortunately, it’s
a pretty fast fix as I do not need to re-run the day’s stats to correct
it. I just grab all the tot_team records for that day, move them to a
temp table, dedupe, and stick them back in. Doesn’t take too long at
all.

That’s what I’m doing now, and expect for it to wrap in the next 15
minutes or so. However, if all goes smoothly that still puts the
5-Nov-1998 stats run starting at 02:00 so stats will definitely be a
couple hours late this evening.

Just bear with me, I’m banging the numbers into shape and we should be
back to normal by morning.

On an entirely unrelated note, I’ve been fiddling with perl lately
and I’ve hacked together what is probably the single-most useless
waste of time I’ve coded in quite some time.

http://sunrise.slacker.com/~nugget/cdinfo.cgi (at home)
http://miata.hfdirect.com/~nugget/cdinfo.cgi (at work)

:: 05-Nov-1998 15:31 (Thursday) ::

Something unhealthy happened during last night’s stats run and many
participants/teams are underreported. I’ve not yet had a chance to look
into it, but as soon as I have information it will be available here.

Thanks for your patience.

(I put in a 15-hour day at hfdirect yesterday, and I’ve not had any
coffee yet today, so my brain is moving a bit slow… just be
patient with me.)

:: 04-Nov-1998 13:00 (Wednesday) ::

D-Oh…
One of the testing boxes for the FBA crashed yesterday.
On a good note, it wasn’t the iMac.

:: 04-Nov-1998 06:19 (Wednesday) ::

Here’s what I should do, if I had some time :

1) Release the new public code (probably somewhere on nodezero)

2) Really understand what’s going on (again) with this @~&#^!
nasm / non-nasm K6 core.

3) Convert the RC5 cores to NASM format.

4) Take a look at the K6-2 performance with the RC5-MMX core.

:: 04-Nov-1998 06:01 (Wednesday) ::

Whoops!

It looks like a /minor/ glitch in the stats run tonight caused the overall
participant offsets table to get wiped prematurely. This has the unfortunate
side effect of not displaying anyone’s overall participant ranking offset
on tonight’s stats.

Shouldn’t happen again, but I’m not quite positive how it happened this time.
I’ll be keeping a close eye on things during tomorrow’s run, for sure. It’s
been a long time since something like this cropped up.

I wouldn’t worry too much, all-in-all things are going pretty smoothly.

Thanks for your patience and understanding.

:: 04-Nov-1998 01:38 (Wednesday) ::

I have a lot of projects that I’m getting ready to completely
finish and make available. Here’s a time frame on some of my active
TODO items.

Coming soon….

By Friday (11/06/98): Public web statistics for www.distributed.net

By Sunday (11/08/98): Finish initiative for stable distributed.net MXes and mail
that absolutely will never bounce.

By Monday (11/09/98): Completely functional subspace plots
at http://www.distributed.net/statistics/graphs/ including
animation over time and a description explaining how to
read the graphs.

By Tuesday (11/10/98): Finish plan to reorganize distributed.net DNS for usability
with monitoring scripts and contact information. This includes
a time frame on the completion of the reorganization.

:: 03-Nov-1998 22:17 (Tuesday) ::

Stuff that is in progress or will be shortly for the MacOS client:

1) Continue to make common source easier to port to MacOS
2) Incorporate new 68k cores
3) Get up to date (.418) on source
4) FBA w/ controller app
5) Remote Cows for Mac
6) Incorporate graphing a la 2.6xxx
7) Tweak UI if necessary
8) Applescript support
9) AltiVec cores

:: 03-Nov-1998 14:35 (Tuesday) ::

A fix for the Mac Faceless Background App (FBA) is now in testing.
It addresses a couple of the issues being reported, namely:
– Crashes on execution, especially on the iMac.
– It now uses the standard config file.

:: 03-Nov-1998 01:38 (Tuesday) ::

Website-update notification scripts are done. Now I need to know how to
implement them: email translators upon a page-change, or use the data to
construct a comprehensive What’s New page.

Additionally, http://www.distributed.net/credits.html has been updated.

I’ve just completed commentary on the mildly famous Halloween Document
posted by Eric S. Raymond. My essay is available at
http://sine.com/creative/oss-response.html .

I’d appreciate any comments on, well, my comments.

:: 01-Nov-1998 23:43 (Sunday) ::

In the past, every time the master finishes assigning a certain subspace,
a master operator is required to generate the new subspaces and set them up
to be distributed. While this wasn’t a big deal 6 months ago when we only
rotated subspaces every couple months, now it is. We’re now rotating
subspaces every 12 or 13 days. This has become a major problem, since it
always seems that the master runs into the end of a subspace in the middle
of the night when all the operators are asleep.

I’ve recently written, tested, and implemented a script which takes
care of this serious problem. Basically, a master operator has to tell
the script which subspace should be assigned next and then when the current
one fills, the script rotates the subspaces and mails all the master
operators.

Now, proxies are less likely to be assigning random blocks because the
master has run out. I’m now working on my #5 project, listed above. More
details as I progress….