Thursday, January 23

A list of topics organized to be vaguely resembling a resume

This is a stream of thought - I am not sure if I want to turn this into a resume or not. I am looking for new challenges - for the "right kind" of opportunity, and for that I need to define what "right kind of opportunity" means for me.

I've since refined some of this into a resume, and am adding references. I expect to continue refining my thoughts about what I want to be doing, and I will also be trying some non-employment mechanisms for sustaining myself.

As an employee of Legi-Slate (a washington post subsidiary, kind of like congressional quarterly - think "what google would have been in the '90s, if you built it out of leftover mainframe technology from the '70s"), I accidentally triggered a meeting of what I was told was the Federal Reserve Board. (Since I never personally met the people involved, I cannot say who actually met, nor provide any other details.)

Basically, I was updating the storage format for saved searches (think: "rss feed"), and a misplaced character caused some searches to be treated as new, and old results were presented as if they were no results. This resulted in the appearance of a massive flood of legislation on banking issues when in fact the legislation was not at all new. No one realized this until after the members of the board had been flown in to the meeting and sat down to review the activity. I was blushing for days when I heard what my mistake had done. Moral of the story: some issues only become visible when you go live.

Also at this point in time, I remember writing Linus Torvalds and coming up with a representation of relevant unicode characters for use in the kernel. I wanted that so that I could represent apl unicode characters in emacs. This was before I gave up on emacs because of RSI. Moral of that story is: do not type too fast,  for too long, with too much keyboard chording. Or, if you do, give yourself a few years to heal, afterwards.

Back when I was active in Debian, I  helped design the package structure (I remember talking people into using makefile format, and documenting it enough so people understood it, for debian build rules). I also helped design the organizational structure and the concern for copyright of debian-legal. And, I was on the technical committee for a time, and talked people into accepting commercial involvement and value. I eventually ducked out because I wanted to be spending more time with my family, and at that point in time, whenever I submitted a package proposal some eager person with more time than myself would have a draft package submitted before I felt comfortable submitting mine. They're doing fine without me, I think (a recurring theme, for me).

Also, back in my debian days, I had some involvement in working with the design and character of voting systems.

Anyways, I'm pretty big on backwards compatibility, simplicity, and zen coding, a concept introduced to me by Marcus Ranum, though I'm honestly still not convinced that the implications of the industry focus on firewalls - something I partially blame Marcus for - has been properly understood and managed. Having mentioned him, though: he's an excellent person to have on your side, if you care about network security.

Also around this time, I was concerned with the restrictions on cryptography in the context of freedom of speech, so I sort of implemented rsa encryption in my email signature, using J. Back then, though, J was not free software, so could not be included in Debian. That issue has since been addressed.

As an aside, I have often felt that Ken Iverson implemented J as a personal favor to me (not strictly true, but I do remember passionately arguing a variety of issues with him, which have shown up in J's implementation). I feel really bad that he died at a time when I was dealing with family issues and ignoring J as a vehicle for education about software issues. That said, the issues surrounding APL and computing history are subtle to convey, and other people have provided far more solid contributions than I.

Year 2000 issues along with internet competition disrupted LegiSlate's business and the company got split up and sold at the last minute when we realized that y2k certification was scheduled to be completed in september 1999 instead of at the time when government agencies needed to sign contracts. I switched from LegiSlate to USAToday at that time, with a significant pay cut, but incredible job flexibility. I was granted permission to work remotely (after an initial shake-out period) and with permission from my manager I was moonlighting another job or two, sometimes during business hours, sometimes after hours. 

Moonlighting jobs included:

* teaching java and javascript (typically in intensive 3 day classes, presented in government agencies or business facilities), this was under the auspices of a company called "Carrig Learning". This ended with the dotcom crash.

* implementing a batch of perl modules to "CPAN Standards" to audit (mostly solaris) computers for the justice department. This was my first serious exposure to perl, and was a lot of fun. This ended with the delivery of the code. (It was suggested that I pick up a government clearance for further work of this nature, but my family background is opposed to too much involvement with military and government issues - I am happy to be of service, and I love those guys, but it's not something I can allow myself to get involved in, personally. Or, at least, that's what I try to tell myself - everything winds up being connected to everything else, nowadays.)

* network operations for a company called aspen research. I also remember drafting a proposal for use of pgp for government security operations. That was fun, also. But I was not willing to shift my schedule with Aspen from evening to daytime, and I was documenting some systems which apparently I was supposed to be ignorant of, so that ended that job.

Meanwhile, for USAToday, I and one other guy walked in to replace the entire net ops group (who I never met) doing 24 hour on call support for their web publishing system. Other fun projects there included:

# "Weather Customization". This allowed a user to enter their zip code and we would put on the weather page a small weather forecast based on the nearest weather station to that post office. The fun thing here was that this was done entirely through static text files, publishing out the full weather forecast every few hours recast as javascript data and chopped into pieces for quick loading. This put something of a strain on the small computers we were using back then and motivated a switch from solaris to windows for our web servers (and, consequently, much more strain on the people involved, but at that time it was cheaper to hire people who claimed experience with windows). This system has since been replaced.

# "Ad Inventory Support". Under the guidance of Kevin Lefew (another really smart guy), we built a system to manage and forecast ad delivery for our site. Briefly: ad sales are based on future web traffic, and the company is faced with a constant struggle between delivering too many ads and too few ads (both of these upset advertisers: while they want their products to be known to the people that have use for them they do not really like upsetting people with too many ads - of course, this is an impossible task to do perfectly, and at times everyone involved gets stressed out. Anyways, the same techniques which are used in trading systems on wall street can be used to better approximate web page delivery. We did this first using C# .Net but it was an early version with a 1GB limit on process memory so we moved most of the computation into MS-SQL of all things. (APL would have been a better fit for the job, but experienced APL programmers get paid too much, which ironically has made the language extremely unpopular.) This system survived a few years, but nowadays this kind of functionality is supported by other firms.

Meanwhile, one advantage of working for a news organization is their anti-deflationary economics. Basically, when things get bad people turn to the news to find out what's going on. The trick, of course, is cashing in on that. This can have a dark side, but can also be a force for economic stability. In the 2008 economic crash, for example, I asked our management to basically offer ads for "free", with our lawyers making arrangements for later compensation. I did not actually participate in those contract negotiations, and do not know what monies were or were not involved, but I think the principal is sound: with the right kind of company, you can generate a lot of value through good faith negotiations.

More generally, advertising has always been a gift economy - not everyone likes it, and you are probably acutely aware of both negative and positive aspects of advertising. Meanwhile, the internet turns higher education and art - mathematics, porn, military information, details of how to prepare and store food, foreign language skills, information about every human activity - into a potential vehicle for advertising.

The trick is: how do we link up the people that want to do things with the people that want those things. How do we keep alive jobs and skills which have largely become redundant while supporting the freedoms which got us here in the first place?

The answer seems to have something to do with the nature of leadership (which involves deep sensitivity to people's needs) and with a deep tolerance for redundancy and dissonance. In a sense, we need to keep solving solved problems, because otherwise we will not remember how to solve them.

Anyways, in my case, it's now time for me to leave Gannett and move on to other challenges. Basically, the company has gotten too rich for the scope of issues which I could tackle within its walls. I think and act strategically (in the sense of the relationships between the sorts of technology I have familiarity with and people and their lives) and I need tougher challenges (or, more precisely: I want to either make my perspectives useful or I want to find out how they need to change).

Possible career directions include:
  • Food distribution in tough neighborhoods (lower income parts of Detroit, maybe)
  • Entertainment (game modding or narrative, maybe)
  • Education (maybe a focus on mixing low tech survival skills with health care and computing)
  • Family (they are incredible people)
  • Optical computing infrastructures (building up computing devices from the technologies behind phone switches)
  • Get a Ph.D. (Bill Yerazunis suggested I get one and apply to MERL).
To my knowledge, there is no one single job where it makes sense to combine all of those directions into a single employment path. Meanwhile, success depends on not wavering from your path.

So my struggle is: do I apply for a strategic position at a large company (Google?) or organization (Civil Service?) or do I go it alone for a time, focusing on some small immediately tangible contributions while neglecting others, get a Ph.D., something else?

There is just so much I want to do, so much I want to learn, and so little time to do it all in.

No comments:

Post a Comment