Hey, the "spy" dude inside google to me the interview notes from my interview. That one dude at the end really didn't like me, and sometimes I don't think they actually wrote down what I said. Still, it's awesome feedback, even if I do feel a little stupid.

Kelly was a strong clusterops manager candidate.  I would like to bring Kelly in
for an in-person interview right away.

Kelly has been at Nvidia for roughly 6 years.  He has had a good career there
and he has learned a lot.  He was one of the core people that brought linux to
Nvidia and got them off of Sun equipment.  Kelly is looking for a new job at a
company like google where he can get exposure to new things and to find some of
the excitement that he feels has been lacking from his Nvidia job lately.  Kelly
enjoys management and he wants to continue to pursue it as a career.

On the technical side I found Kelly to be pretty adept.  I did not spend too
much time on this area as others will be able to probe deeper when Kelly comes
in-person.  Kelly suggested dmesg, top, and /proc/meminfo as places to find the
amount of ram in a system.  Kelly knew that resolv.conf held the ip address of
your dns server.  Kelly explained that the reason a machine might have an hda
and hdc, but skip hdb is because you have two master drives.  He understood that
you might want two masters for performance issues.  Kelly understood the
difference between a switch and a hub, correctly noting that switches use mac
addresses to make decisions.  Kelly was unable to recite the 3way handshake (he
knew there were syns and acks, but did not know the exact sequence).  Kelly
easily explained what a syn flood attack was.  He suggested shortening the tcp
time wait as a technique to defend against syn floods.  Kelly had never heard of
syn cookies.  I told him that he might want to look them up, so its a good
followup question in further interviews.  Kelly was rusty with carving up
subnets, but with some hand holding he was able to make the calculations.  Kelly
correctly calculated the number of hosts in a /24 and in a /28 and he was able
to convert 101 to 5.

I asked Kelly what it takes to be a good manager.  He stated:

1.  be impartial, look at all sides of the issue
2.  keep a level head, don't react impuslively
3.  do all you can to keep people motivated and happy and help them to define
and achieve their career path

This is a great answer.  I especially like number 3, which I feel is very
important!

I asked Kelly what it is about management that gives him the most satisfaction. 
He replied:

1.  loves molding jr. people.  He likes to be a leader for them and to
facilitate their learning and achievements
2.  likes to stick up for his team.  He feels great satisfaction in being able
to advocate for the members of his team

I feel this was a great answer.  Once again, Kelly focuses on the individuals
that he manages.  Poor answers to this question are when the candidate gets
satisfaction about his/her own achievements rather than those of his teammates.

Kelly is a big fan of google and I think he "gets" what we are all about. Google
is his homepage and Kelly says that he loves google because it is simple, fast,
uncluttered, and stays pure to focusing on giving the user what they want. 
Great answer!

I gave Kelly the opportunity to ask me some questions and he had a long supply
of them!  He was very curious about how google worked and I think this was a
good sign.

I feel that Kelly had a solid technical background, a solid managerial
philosophy, and a properly aggressive and positive attitude that would make him
a successful manager at google.  Lets bring him in for an in-person interview so
he can talk to other people.

When you came in for your 1st round, you met w/ Marc Felton, Sam Mackness, and Keith Kleiner. Ive pasted their feedback in that order:
 

Overall Conclusions:
I read the the phone screen mentioned syn cookies and told him he
"might want to look them up" so I asked him if he knew what syn
cookies were. turns out, he didn't want to look them up, but thought
they were very groovy when I explained them to him.
 
asked him what he knew about networking, since probably clusterops has
to deal with that kind of thing from time to time, and he said it was
very low in his list of areas of interest, and basic networking was
already covered by Keith. He drew me a network he set up once to load
balance work across 100 diskless clients, and he used 4 NIC cards and
he could tell the clients what their gateway was and get traffic in
that direction balanced but in the other direction it would all go
down one link. 
 
so I went to troubleshooting, office subnet is 10.0.0.0/24 and someone
calls him telling him the network is down
-tries his own box's network (works fine)
-tries to give the user an IP address to browse to in case its DNS
(cant browse to the IP either)
-tries to go the user's machine (fine) checks the arp table (see wierd
192.168.0.1 address)
-run ifconfig (yup, 192.168.0.4) -ok someone running DHCP
-get Mac, look on switch, trace patch back to desk, "take a rolled up
newspaper and" (makes whacking motion as clicks tongue, but he was kidding)
 
asked the ns1 goes down problem, he put the IP alias on ns2 in 10
seconds flat.
 
asked the re-ip many machines on mars question
-before make changes, put a script in place to put old network back
be more specific about how you'd do that
-use files in /etc/sysconfig/network-scripts/ifcfg- if can't ping
router, switch back
-put a longer term cron job in place that you can take out later that
reverts one more time if you dont trust the network (part of the way I
ask the question is to point out that you can't be sure the networking
in the new subnet will work and you can't lose any machines and you
have no console)
 
talked a little about perf at the very end, he has had to let someone
go, took 4 months, one of the first things he did as a manager, he was
complaining to his boss that the other half of the 2 person team
wasn't pulling his weight, so the boss made him a manager and told him
the problem was now his to solve. lets make sure we dont hire whoever
his manager is, but hiring Kelly is a good idea.
 

Sam Mackness:


[What do you currently do at Nvidia?]

I manage Engineering IT reporting to the VP of Engineering. There is also a
Corporate IT Manager and a Network IT Manager. My team is currently about 10 sys
admins. I spend about 90% on my time on pure management/technical design. I used
to spend lots of time writing Perl and I love hopping back into that every so
often but its not a specialty of mine any more. I definitely cant be the
technical bottleneck on anything and I need to spend my time where its needed. 


[Why are you a manager?]

You can only do so much by yourself and if you gather a team together and run it
well you can accomplish more powerful and complex things. An example of this is
taking over the perforce design and administration at Nvidia. I wanted to make
an impact from IT and change the way the business works for the better. 
 

[Tell me about a recent technical challenge you faced?]

[Here Kelly went into a pretty long explanation of a pretty common problem faced
by growing companies; Netapp/NFS limitations. He talked about maxing out
concurrent NFS requests and having to rearchitect a solution which included a
big Hitachi datastore and multiple frontends running some Veritas software.
Kelly actually didnt like this solution because he thought it was too brittle
(eg. he didnt like the proprietary Veritas software) and he actually
recommended against it. It was clear that something needed to be done fast
however and he put aside his personal bias and implemented it. He said that the
new system ran fine for 30months, and he has just finished replacing it with
something else (I didnt ask what because wed already wasted lots of time on
this question). 

 

[How do you grow a team/manage its growth?]

Well I can only talk about what Ive done before and at Nvidia weve had success
hiring contractors. Ive always wanted to get a feel for people in an actual
work environment, plus you get people in the door faster.

 

[But this comes at a cost right?]

True, but its worked well for us.[Boo]

 

[How do you facilitate career growth among your reports?]

Sit down and find out where the person wants to go in their career. Where do
their interests lie and get them to figure out what might be next for them. Make
sure the more senior guys keep an eye on the junior ones and try to give out
challenging projects to those than want them. You also have to set clear goals
and milestones for more junior people. They have to have a clear focus and
consistently reach their goals. Its also good to let people fail sometimes,
its a good learning experience. [Good!]

 

[What is your teams dynamic like?]

Everything is very open. People know who is good a what and that they can
approach anyone with a question. My role really is to facilitate this. My group
is also like a gang, they all stick together. Im a firm believer that the
manager is responsible for the attitude of the group.

 

[Is anyone problematic?]

One person has a hard time communicating to the customer, but weve had some
good talks recently and he understands that its something he has to work on.
Hes very good technically so I dont want to give up on him.

 

Verbal Communication Skills/ Google Culture Fit, Overall Conclusions:

 

Kelly seemed pretty Googly. He was well spoken, clear, and appeared to be pretty
sharp. Hes another one of these candidates however that is here interviewing
here for not the best reasons (he is now perfectly happy at Nvidia) and is
really not that excited about working for Google and Im certain that he doesnt
really get what were all about. But having said this  judging purely on his
responses hes a solid IT manager and I dont have any real doubts that he could
get the job done. Hes not however in my book a superstar.

 

Keith Kleiner:

 

Purely looking at what he has to offer as a technologist and as a manager, my
in-person interview confirms the positive assessment of my phone interview. 
Kelly's skills were offset, however, by his lack of any excitement whatsoever
about google.  Kelly is happy in his current position at Nvidia and did not make
the overtures to me during the interview that would suggest he is strongly
interested in google.
 
Kelly and I discussed a number of management related issues and he said all the
right things.  We discussed how to deal with an adjacent team to yours that does
not live up to your teams' expectations in terms of responsiveness, deadlines,
etc.  Kelly said that he had actually dealt with this exact issue at Nvidia and
his explanation of how he handled it was perfect in my opinion.  Kelly said that
the first thing he did was have a long talk with his team to get it into their
heads that the opposing team was not an adversary.  He had to tell his team to
treat the other team with the utmost respect and to give them the best service
possible, even if it was not recipricated.  Kelly did not force this way of
thinking onto is team, but rather reasoned it out with them over the course of a
long meeting.  He says he felt good about working with his team on this and that
at the end of the meeting they agreed that being positive was the proper
approach.  Meanwhile, Kelly would work with the manager of the opposing team or
his own manager as appropriate to get the opposing team to improve its ways
through management channels.  Kelly makes sure his team feels comfortable
talking to him and he wants them to be able to vent to him.  Better they vent to
him than to the customer he says.  I especially like how Kelly calls other teams
and people within the company customers.  This is the way I like to operate as
well.  Using the word customer signifies to me the positive type of attitude and
approach Kelly would use at google when working with others.
 
Although Kelly did well on the core discussion above, he did not show me the
strong interest in google that I like to see in candidates.  I gave Kelly the
opportunity to ask me questions about google culture, etc so that he could learn
more about us and hopefully get excited.  I specifically asked him what would it
take to get him interested enough in google to leave Nvidia and work for us.  He
responded by asking me various questions about our budgeting processes,
management style, etc.  All good questions but not lined with the enthusiasm and
curiousity that I was hoping for.
 
Overall, I am supportive of Kelly for hire at google.  My assessment leads me to
believe that he would be a good manager.  Unfortunately, his lack of enthusiasm
and interest dampens my impression of him so I can only give him a 3.1.  I think
we should interview him further and hopefully as he talks to other people his
interest level will heighten.
 


Marc Alvidrez:

 

He had never managed a distributed team, but had been a part of them.

The primary challenges he anticiapted with a distributed team were
timezone, the remote people not feeling part of the team, and the remote
people having a sense of cultural connectedness.  That is, depending on
where they are they may have a different sense of a team, method of
escellating problems, work ethic, jargon, etc.  I asked how he would
address these challenges.  He said that he would want to visit the remote
sites frequently (how frequently, he didn't know).  He would attempt to
treat everyone as a part of one team.  But he was at a bit of a loss for
how to deal with a remote person who is underperforming.  He got stuck
thinking about how coaching could work across different timezones.

 

I asked him if he had ever fired a person for cause.  He said he had done
it twice and that, in each case, they had really earned it.  He had worked
closely with them to understand why they were failing.  He made a clear
distinction between failures of ability, training, etc., vs. failures of
motivation.  The first category have easy solutions, and the second
category doesn't.  He stated that he "wanted to sve these people," but
couldn't do it because it wasn't fair to the other team members or to the
company.

 

Technical Ability:


I asked him about some perl scripts he had written to aggregate real-time
statistics of a compute cluster and he showed a good understanding of his
particular implementation and its limitations.  He was especially quick to
mention the scaling problems with his design and implementation and could
discuss how he would address the limitations if he had it to do again.

 

I asked him how he would copy a directory tree from one place in a
filesystem to another.  He said he would use 'tar' and came up with this
command line:
 

 # tar -cpf - | ( cd dest ; tar xvfp - )

 

I asked him to debug a stuck popd process.  He followed a logical
debugging path that led him from 'telnet' to 'strace/truss/ptrace'.  He
did a good job here.

 

Technically, Kelly has his wits about him.  His solid operational
background and experience shines through even if he doesn't have all the
details at the tips of his fingers any more.

 
 
Notes:
 
Overall Conclusions:

 

I was not blown away.  I don't think that there was anything in the
interview to absolutely disqualify him, but, then again, nothing made me
want to hire him immediately.  I think he probably could do a good job,
but I'm not sure he's a peer of the other managers at his level in the
company.
 

 

Cody Smith:

 

What has been the hardest part about ramping up the size of your clusters?

What tools do you have for tracking your clusters and organizing repairs?

What is a syn cookie?

What trouble-shooting steps would you go through if your pop3 daemon was stuck?

If you have an 80GB disk that appears to contain two 60GB files, why?

What challenges would you face with a geographically distributed team?

How would you handle an employee who is having problems?

 
 
Notes:
 
Shell/Perl Scripting:

 

Didn't any any specifics, but talked to Kelly about a perl agent that would
track proc usage on 100s of machines and feed them back into one central source.
 He admitted without us pointing out that his solution wouldn't scale well.

 

Unix SysAdmin Knowledge (Basic/ Advanced):

 
In the question of having one 80GB disk with two 60GB files on it, he said it
may be a hard link, the disk may be corrupt, the files might be sparse.  These
are great answers.

 

If your pop3 daemon is stuck, first telnet to 110 and see what comes back.  If
nothing comes back connect to the machine and see if the process is running.  If
so check permissions to see if pop3 can see the mail spool.  Check the logs,
check to make sure there is enough disk space, run strace -p to see what it's
doing.  This answer was also fairly complete.

 

Linux experience (User Level/ SysAdmin Level):

Experience with Large Scale Systems & High Availability:

 
He discussed at length the current setup of the clusters at nVidia.  They use
ganglia for monitoring now, and lsf.  Their repairs team consists of one person.
 Apparently they don't do repairs enough to have automated diagnosis.

 

He discussed setting netboot for diskless machines under the old infrastructure.
 He says they're now using PXE.

 

Networking Knowledge:

 
We asked about syn cookies, because Keith made a point of stating that he had
missed the question about syn cookies in his phone interview.  He gave a good
overview description ("it sends back the state to the client, so it doesn't have
to keep track of it"), but he couldn't describe exactly how it worked.  Given
that Keith had suggested he learn about syn cookies, I was a bit put off that he
hadn't done his homework.

 

Problem Solving Ability:

Experience Relevant to Google/ Knowledge of Google:

Verbal Communication Skills/ Google Culture Fit:

 

He communication skills were excellent.  He was very clear and well spoken.  I
see no reason why he wouldn't fit in fine at Google.

 

Leadership:

 

We asked about managing a geographically distributed team, and he said: he said
he would anticipate the following problems: time differences, people in remote
locations would feel like they're not part of the team and the team culture, he
gets no face time with the remote employees.  He said he would prefer to split
up teams that are seperated by some distance, which was interesting.  He also
said that keeping track of remote employees would be easier, because all you
have to look at is their results.  This is a bit off in the case of clusterops. 
He also said that coaching remote employees is hard because you can't relate as
well using video-conferencing as you could in person, so he said he would travel
to remote locations periodically, if possible.

 
With problematic employees, he gave a long answer, but it basically came down
to: determine the root cause of the problem and fight against that, don't fight
against the symptoms.  This is really a superb answer.

 

Overall Conclusions:

 

I have a few reservations, but I think Kelly would be a great clusterops
manager.
 

 

Rudy Winnacker:

 

Kelly was asked how he would rate himself on a scale from 1 - 10 as a perl coder
and he stated 5. He could not say what 'use strict' does, and didn't know -T. He
said that taint checking could be useful to prevent people from "hacking" or
"buffer overflows". He did not know whether/how DNS uses UDP/TCP. He guessed
that a website such as cnn.com would use more outbound bandwidth than inbound. 
 
When described a situation where his team is receiving complaints about DNS
abuse from a specific cell, he first gave a very short answer in terms of
stopping the job. After some prodding, he started thinking around the various
issues that might be involved, including setting up dedicated nameservers, load
balancing them, using dnscache. When asked how he would plan for the DNS needs
of a cell he stated several times that he would observe the behavior of the cell
versus DNS when running real jobs. At the end of this line of questioning he
proposed replaying the DNS requests from a job in a test environment.
 
When asked what skills he would expect from an SA1 he said a strong general
background with computer systems, and not any specific skills. The concepts are
more important and skills can be taught. Knowing vi would be nice but not a
requirement. Supposing he had too many applicants that would fit these
requirements, he would then prefer someone with UNIX experience. They would need
to understand IP addresses but without having to know how many hosts are in a
/29. They would not need to know tcpdump.
 
From an SA2 he would expect direct experience with the requirements of the job,
and that they know 'everything about the OS'. Asked to expand on this he
mentioned: patching, rebuilding (in general), disk i/o, troubleshooting,
monitoring, variety of h/w platforms, and as an example, knowing that a one-disk
machine is not ideal for a dbserver. Asked how he would know the right
candidate, he said he would bring them in on a contract basis for a probationary
period. He would ask them to describe a project they had completed, since the
better candidates will enthusiastically dig into answering this question.
 
When asked his most difficult technical experience/coolest hack he described
setting up a diskless linux cluster and balancing network traffic through two
routers by hard coding routes on the hosts.
 
When asked about the professional development of a system admin he stated that
junior system admins would be provided the base skills required with lots of
direction. Intermediate system admins would start architecting their own
solutions, combining elements of their knowledge, or else start getting into
scripting/automation (but not both because that doesn't work). Senior system
admins would either work on more complex project designs or else
scripting/automation projects, requiring the help of other sysadmins. Beyond
this, one would then move away from the technical; he's never seen anyone stay
technical advancing past senior system admin.
 
Why did he go into system administration, then management? He wanted to know how
computer system work, not just how to use them to compile programs. He got into
management originally because that created the opportunity for him to accomplish
the projects he was interested in at the time but which had ben blocking on
management approval. He likes management because he likes to help mold the
junior system admins; he would provide them mentorship by having more senior
system admins assist the more junior ones.
 
Summary: I was not very excited by this candidate. I expected more technical
abilities to have come out of the interview and although we visited the issue of
the techncial role of a manager several times he made it clear that he didn't
think a good manager could be technical and also do other things such as keeping
the workplace fun, doing performance reviews and represent the group to the rest
of the organization.
 
I was impressed by some remarks of his indicating that he sees the manager as
having certain responsibilities to the people he manages and thought that would
be a good fit.
 
But I was confused by his answers to the different ways we asked how he could
measure the success of his team without being technical at all. His answers came
down to delegating that to more senior sysadmins, but I found this confusing
since he would measure the competence of the more senior sysadmins by how
enthusiastically they described their projects. He also never made it clear how
he came to his view of management as necessarily non-technical while he himself
started in management because he saw it as a good way to get important projects
approved and finished.
 
I may be wrong about this but I believe the clusterops manager should have at
least a little more technical ability and interest in keeping sharp technically.
I was concerned by the lack of discussion of objective metrics to gauge his own
success and the ability of system admins. I didn't understand his big picture of
the career of a system admin - it seemed that in his view at some point system
admins just stop doing system administration and start managing, while
forgetting almost all of what they had learned. I did not get the feeling that
he would help grow the team with "absolutely the best" system admins (from the
job description), or how he would maintain the confidence of his team that he
would help them to further their professional careers while also performing
vital services for the company.
 

Todd Curtiss:

 

 
 
Shell/Perl Scripting:

  Self-rate 1-10: C: 1, Perl: 5

 

  Kelly commented about not being able to code on the whiteboard before I got to
ask, and made frequent reference to not having done hands-on technical work in a
while, so I took this as an indicator not to bother with a scripting exercise.

 

  Why declare 'use strict' in perl?  No idea.

  Why use -T, "taint" perl?  No idea.

  Why would you want to restrict input or parameters passed into a script, maybe

a cgi script? "For hacking purposes, to prevent a buffer overflow or something".

Would have liked some more depth here.

 

Unix SysAdmin Knowledge (Basic/ Advanced):

Linux experience (User Level/ SysAdmin Level):

  Maybe mid-level? I didn't probe deeply on SA-specific skills, but elicited
from our conversation that Kelly understands some of the important underlying
concepts (how network boot happens, for example), but hasn't ever attained real
expertise in any of them.

 

  Does DNS use TCP or UDP?  Don't know.

  How would you reduce impact of DNS lookups/volume?

  Caching DNS, or enter hosts into hosts table on local system.

    Decent, but no further depth provided when I asked.

 

  How do you know when you need additional DNS servers?

  I was trying to elicit some capacity planning methodology here, and the most
specific answer I got was "measure queries under load". The best part of Kelly's
response was that you might be able to "replay queries from actual traffic". I
asked what kind of load, how to measure, what metrics to use/collect, how to
prevent overload (trending), but didn't get any of this. 

 

  Most websites: more I/B vs. O/B traffic?

  O/B -- most cable modems do faster download, and most sites send back more
info than they take in.

    OK, but was hoping for discussion of HTTP specifics. No biggie though.

 

Experience with Large Scale Systems & High Availability:

  Has managed a 3,000 node Linux cluster (LSF) for jobs running at nVidia.
Handles management of internal tools (Perforce, etc.).

 

Networking Knowledge:

  Rudy asked Kelly if he'd be able to calculate hosts in or netmask for a /29.

No.

 

Problem Solving Ability:

  Hard to gauge. I asked a few questions to try to elicit how Kelly would solve
a service problem, but he repeatedly fell on his "rustiness". When I asked him
about his "hardest bug", he didn't really have an answer. I then reframed this
as "coolest hack", and he provided details of the primary project he's done at
nVidia, in converting to Linux cluster-based compute farm. While this was
interesting, he didn't really highlight innovative methods nor did he convey why
he chose to solve the problem the way he did.

 

Experience Relevant to Google/ Knowledge of Google:

  Has led teams of SA and tools folks at nVidia for under two years, his only
management experience so far. Previously was a hands-on SA/NetworkAdmin for
nVidia, BSDi, and others. Surprisingly lacking some of the Unix and Networking
skill I'd expect from someone only two years out of doing this full-time.

 

Verbal Communication Skills/ Google Culture Fit:

  Good speaker, made solid eye contact. Kelly's expectations of what a manager's
role is in leading a technical group -- as a non-technical contributor, more of
a personnel manager -- doesn't seem to jive well with Googleyness.

 

Management Skills:

  Limited -- only has ~2 years of management experience, leading a small team of
SAs.

  Asked about hiring and retaining a team, he focused a lot on hiring
junior-level SAs and giving them basic tasks to start -- as a proof of skills,
and in a C-to-H scheme (which we don't do here, but I've seen work elsewhere). I
did notice that Kelly's understanding of various levels of SysAdmin were very
squishy, and didn't sync well with typical industry rubric (SAGE, etc.). He said
that he has not yet "found a good indicator of what makes a good SA", and that
it's "a guessing game, believe it or not" to figure out if someone will work out
as an employee. I thought this showed a lack of experience and maturity in
hiring teams.

  Talked a lot about "architecting things", but lacked supporting details and
fundamentals. This was also demonstrated by Kelly in his unwillingness to
attempt a few technical questions I posed. He chalked much of this up to
"rustiness", even stating flat out that "managers can't be too technical" and
that as a technical contributor that you can only expand your influence to a
limited point. I asked how one stays 'technical enough' as a manager of highly
technical folks, and he said, direct quote: "you can only get so big being a
technical problem solver". He gets most of his direction from his "senior guys"
and states that it's important for him not to be in the critical path during
problem solving. He sees this involvement as "failing as a manager". However, he
also states that he went into management from being a SysAdmin because "you
can't influence decisions without being a manager". Worrisome.

 

Overall Conclusions:

  I don't see Kelly managing at Google, and especially not managing highly
technical individuals. He seems to think of the balance between management and
hands-on contribution as a mutually exclusive division, which seems largely
rooted in a lack of management experience. Compared to other candidates with
whom I've talked for the ClusterOps Manager role, Kelly falls far short.

  Recommend a pass. 
 

Don says that D-link networking gear is crap too, he's had bad experiences with it. I can't imagine they have many repeat customers; I'll not be one. I did use their domain blocker to kill "tribalfusion.com" so the woman won't get all the pop ups. Firefox can't seem to block those, although the new "fasterfox" extension I just installed says it knows how to get applet embedded pop ups. Hopefully that works.

Anyway, Friday night the woman and I saw Serenity as planned with the pumpkin patch run. I liked it, she tolerated it. Don says that the show was good so I'll borrow the DVDs from Steve at work and give them a view.

Today I had to go to work, we had a major power outage that took 6 or so hours to recover from. It actually took longer than that, but I left one of my guys there dealing with the slow booting suns while Don and I went to the range. Abusing my managerial powers for once. Anyway, I didn't shoot very well but I narrowed the loads from 4 down to 2 that I think are good. I'll load up more of those two and see what happens next time. We had pizza from the good place for dinner which is always a bonus.

While at work I went to my parent's new place at lunch. Their little trailer in a ghetto neighborhood in San Jose is actually pretty nice. I'll have to take some pictures of it at some point. They really don't have much closet space in that thing though. Their townhouse had it's first open house this weekend, my guess is that they'll have offers by Tuesday.

Wednesday night the woman took me to the city to see "Bauhaus" in concert. They're some 1983 goth band that broke up and formed "Love and Rockets" and then came back together at some point (1997?) and then quit touring and then started touring again, or something like that. She was very excited to see them, I just sat there looking at the weirdo goth chicks. Since I put up with her penguin movie and this concert, we're going to go see Serenity tonight after we visit the pumpkin patch. Some of the guys at work say it's good.

Back from Costa Rica, and of course pics are on the main page. It was a good trip but I'm glad to be back home. I'll admit I'm a wimp, I like having hot water in my shower. :)

I should write more, but I'm kinda hungry at the moment. The woman had to go to work today and I'm home alone doing chores. The laundry is amazingly stinky, things just do not dry in the rain forest. At all. Even the fancy microfiber dry in 2 hours stuff will stay wet for days. It was hell on the electronics too, the camera was fogged half the time and two of the other people's cameras completely crapped out. I also didn't take the camera on the kayak trip because I didn't have any sort of real dry bag to protect it. I figured that with all the duct tape we used to hold the double kayak together that we were sure to sink at some point (it actually held and I got some pics of it).

That sucked, but I finally got all 3 computers back on the wireless network. It took 3 trips to Fry's and two passes though the return line (yea, some god somewhere hates me). The damn Dell XP system was the killer, every card that went in there caused some sort of funky conflict and would lock the system solid after a few minutes/seconds. I gave up on the PCI add-in card and just got a cheesy USB dongle again. It's "only" the 54mbit version and it's only getting 36mbit or something from the 3rd floor to the garage but the woman doesn't care and she likes the blue light on it. I must say that "D-link" products suck major ass, and "Netgear" stuff really isn't all that bad. Netgear has spiffy apps and work on the first try, d-link has XP looking apps (ie big stupid buttons aimed at the 3-year old market) and the crap takes 50 reboots to get it right. Laugh at Netgear if you must, but it works.

I imagine there is some irony in the fact that the d-link router just dropped my connection for a moment and I had to recover the vi session.

Playing with the browser I just read about the Cancun tourists that are trapped (like 40000 of them). That really sucks, I've flown in/out of Cancun and it was a disaster even when things were normal. Bummer about the beaching being toasted too. There is some cow (sorry, "woman") blaming Bush for the troubles. Uh, how exactly did Bush cause this? Man, Americans really are stupid.

Now I'm trying my win98 desktop and it's still working. Very exciting. It's got really good signal compared to the USB widget across the room. Still drops connection from time to time. The router people say that's due to bad signal but their apps show I've got 97% signal. Maybe someone is microwaving a burrito and it's causing problems. I probably should upgrade from WEP to whatever the other types of encryption they're offering, I think all the apps/cards I've got do that now.

Amazon will let me order books that aren't out yet and ship them when they're published. That's kinda cool. What's not cool is their shipping. Shipping a few books costs almost as much as the books themselves. I'm going to have to hit the bookstore tomorrow and see if they're in. I would rather pay 10% sales tax instead of 50% "shipping". Guess you gotta order a lot and get the free shipping to make it work. I tried, but I only want 4 books and two aren't out yet. Can't get the free shipping on vaporware. I also see they can let you download some books. That's kinda cool, but I read in bed and I don't want to have to print stuff and/or have the laptop around when I'm trying to sleep. Good idea for technical reference books maybe, or people addicted to reading from PDAs.

I slept for 12+ hours and it really helped. I can sorta smell stuff still which is nice. Having not moved from the sofa for a while I've watched the backlog of stuff on the replay and am now watching daytime TV. Pretty much it's a series of ads trying to get old people to abuse Medicare for a variety of potions, pills, scooters and who knows what else. It's also a bunch of baby food commercials. I had no idea that baby formula was such a big business. Who knew there were so many woman who can't produce breast milk? The world is just a weird place.

Why is the Brady Bunch playing at 6pm on the WB? I'm sure it was always this cheesy. Oh, and the other thing I've noticed about daytime TV. The image of women they're selling now is sorta gross. Sure, I'm watching elimidate and whatnot but the chicks are just nasty. They're all distorted, wear too much makeup, etc. I'm not just getting old, they're changing. I'm glad I'm not 20. There was a diet ad where some chick was so glad she went from a size 10 down to a size 4. I know the camera "adds 15 pounds" but this chick was dangerously skinny on TV so I can only imagine what she's like in real life. She looked better as a size 10.

Home sick today after going in for the morning manager's meeting. What a waste of time, I should have just stayed home and slept.

I got a new wireless router and it works well, except that none of the 11b devices actually talk to it even though it claims full 11b compatibility. It only works with the 11g card I bought with it. I hate it when vendors don't follow the fucking standards. Do they all take their lead from MSFT?

The wireless part of the wireless router is dead. Like stone dead. It continues to work on the 4 port switch in back though which is good. I'm down in the garage with the laptop trying to debug the thing and man is the net connection fast without the wireless. Could the wireless latency be that bad? Or is my 4 year old wireless gear just crusty? May have to go to Fry's and get the $49 G-router doohickey.

Pictures from the Shasta run on Sunday are uploading now to the main page. As you can see, we have a slab and a well and a tank. Next comes the shed. We also found where the neighbors cleared a path to the edge of the ravine where the creek is. It's way down there...

The crappy neighbors moved their gate back as they said they would, so that's a done deal. The did put up a lame "Metallica Blvd" sign. Guess we know their mentality now. Lame.

They pushed the work from this weekend back again so I actually got today off. Nice. I went to the range with the new rifle and tested the 5 loads I made up. I think that if I can control the fliers many of these will be good. The worst was group #3 which was those fancy black winchester bullets I liked the look of. So much for that. The best load was #5 (on the "S" labeled target) which I found on the net and didn't really think would work. The H414 powder isn't rated to be one of the better for .308 but it sure does work. For those without a micrometer handy, the best load was 0.446" center-to-center. That's pretty good at 100 yards for a non-bench-rest non-fancy rifle. Don's should do better.

Tomorrow we're headed up to the land to check out the concrete pad and the pressure tank and make some measurements. We also got a fancy tarp to cover everything with until we get the shed built. It's only going to be a few weeks but the woman is worried that animals we get into stuff. Animals? I guess they'll be fooled by the stealth tarp technology?

What fun. I was out of the office on Tuesday working at the DR site when the shit hit the fan (again) here with respect to perforce. I was out in general yesterday (comp time for all the weekend work) and pretty much ignored everything. It was nice, and the world didn't end. Sure, more things are screwed up than if I had been here to keep people on task and focused, but we're all still alive. If only anyone here put any value in the ability of one person to keep other people on task and doing the "right" thing. Oh well, not my problem.

I was playing around with the micrometer again and found that ammo coming out of my cheap Lee hand press varies from round to round up to 0.008". Sure, 8 thousandths doesn't sound like much but the whole point of crazy precise brass, high end primers, bullets and powders is to make everything consistent. All that goes to pot if the seating depth keeps changing. I gotta get the conversion for the other press to allow it to use smaller calibers than the 50 and start using that.

Speaking of the 50, I got my registration paperwork back. The DOJ took it upon them selves to re-register my AR to my Milpitas address too. I was wondering how I would do that, and now it's done. Creepy, but a nice touch. I'm just wondering when the local PD will knock on my door and have a chat?

Also yesterday I had a refresher horse lesson. The main trainers at the barn the woman goes to (went to?) left and took all their stuff with them. There are a few horses left, but a lot of the tack is missing so it took almost an hour to prep for the lesson. That sucked. I did OK, didn't die even though the horse tried to buck me off, etc. Apparently everything I did when the horse flipped out was wrong and only making things worse. When he reared up I fell back into dirt bike mode and tried to stay on top of him and cranked down on the reigns. oops. No one told me that.

Man I hate my fucking job. Maybe it's not really the job, it's just the people at it.

Working all weekend is fun. OK, it's not really. We worked all day Saturday, and Sunday morning I had to go to Napa with the woman to meet the wedding officiant, and then I came back and went to work again. We got 95% of the farm converted and all that's left is a few of the junk PIII system's that can't auto-boot from the net. Not bad at all.

Today I get in to work and find our shit ass Wipro contractors up to their old tricks again. Not only do they want us to supply them with hardware, but they want us to set them up and configure the environment for them too. So we're outsourcing to them, yet we have to provide equipment and setup and our engineers have to help their engineers do the work because the Wipro guys are too stupid to actually work with the data and stuff that we've given them. How exactly are we saving money by supplying, training and basically doing the work for them? I'm sure some VP gets a blow job out of all this.