pyesetz: (woof)

Suppose you need to convince a client to buy your company’s product, so you decide to take him on a hunting trip — and you bring along your dog.  While jawboning to the client about how great the product is, in order to emphasize your company's ability to complete its projects, you shoot a duck out of the sky.  The duck lands in a marsh, so your dog goes and fetches it for you.  As you take the hunk of meat from the dog, you give him a small prepackaged treat and think, “What a stupid trade this dog just agreed to.”  But who is really getting the better deal here?

Suppose instead that you fail to shoot any ducks, so your dog has nothing to do — yet he still gets his dinner when you go home, just like every day.  The dog gets heating and air conditioning, food, vet care for his boo-boos, and an appreciative boss.  He gets fed regardless of whether you end up pulling off that deal with the client.  He got no worries, hakuna matata!

* * * * *

In other news, I gave another lecture at that monthly programmers’ meet-up, which is sponsored by a company that I guess I’ll refer to as “ℙ” on this blog.  I talked about my never-completed doctoral thesis and how it relates to my difficult-but-eventually-completed move to Canada.  (I skipped over the part about how the USA is not actually a “free country” because previous meet-ups clearly indicated that these Canadians didn’t want to hear such talk about our neighbour, friend, and ally The States.)  I talked about the professor that I had hoped would supervise the dissertation and how I had designed the program to match up with his personal proclivities.  I showed some code and discussed how it connected to certain foundational theorems of computer science.

After my speech, a fellow I had never met before, who apparently does not work at Company ℙ, asked me if I was a professor at the local university.  “No,” I replied, “I just sound like one.”  He asked to see my résumé, so I showed it to him.  He was apparently not expecting to see that I have spent the last seven years doing web-monkey work at Company 𝔾, so he never did talk about whatever job he had wanted to offer — which is too bad because the fellow seemed to be quite well off and I could sure use some dough.

The Company ℙ manager asked many questions about my project, but continued to avoid saying anything about possible employment.  My impression was that my presentation had convinced him that I was not a suitable candidate for his own part of the company (perhaps to be called here?) because his group is all about the “awesome user experience” and my program clearly demonstrates that my visual-design skills are not “awesome”.  Damn it, I’m a content guy, not a pretty interface guy!  But apparently there are other positions accessible through the Company ℙ network, so it still seems worthwhile to go back next month.

The Company ℙ guy who’s big on Haskell wasn’t there this month.  In fact, there were less than a dozen people in attendance because so many people were on vacation for August.  But one guy announced that it was his first day on the payroll so he had brought free beer for everyone!  So that was nice.

* * * * *

Doesn’t anyone need a doggie to go fetch a program for them from the marsh after they’ve convinced a client to buy it?  I can fetch really meaty programs and I don’t need especially-fancy treats for them!  And I can and have fetched from areas of the marsh that most doggies wouldn't dare enter.

pyesetz: (Default)
Ever since I wrote this post, I've had this idea that there would be an "academic experience" journal someplace where I could publish it.  Recently a little bird told me ("caw! caw!") that the Chronicle of Higher Education publishes stuff like this in their "Opinions" section.  So I sent them a copy last Monday.

Despite the short URL, suggesting early web-adoption, the Chronicle seems to have no policy for emailing submissions.  They want double-spaced printouts.  Also they state that they will send a postcard acknowledging receipt, followed by final determination within a month.  (This is much better than Pascal News, which said nothing while sitting on my submissions for a year before finally publishing them—and then only because a change of editor caused them to start scraping the bottom of their slushpile barrel.)

Last week I managed to avoid worrying about it too much, but this week was pretty bad, checking the mail every day looking for a postcard.  Would they actually send me one, or would they just throw my submission away as "not serious"?  When I was at UMass, I applied as a graduate student to the Linguistics department.  After a month, I asked a professor about the status of my app.  He said they had taken no action because they thought the app was "not serious".  Um, why would I pay an application fee for an app that was not serious?  So I applied to the Computer Science department.  It eventually turned out that half the department thought I would be a great candidate while the other half wanted me kicked out of town on my ass.  Yeah yeah, "abrasive personality", yadda yadda.  So I applied to the Slavic Languages department.  They accepted me, although I had told them up front I had no intention of completing the doctorate (which would have required two years' study of German).  So the app that actually *wasn't serious* was the successful one!

There was much to worry about.  The Chronicle pays $300 for accepted submissions, which makes them considerably more liable for what I say than a freebie journal would be.  My post says Rosenblatt's death was a suicide (this is a contested issue).  It says Andrew Wiles' wife wasn't as supportive of him as she claims.  It says David Rumelhart is "a poor scientist".  These claims are not well-supported.  The affected people could sue for libel.

I put my email address on the submission letter as an afterthought, in case they wanted to talk about the libel issues.  Today I got an email from them: "We appreciate your submitting an article to The Chronicle and regret that we are unable to publish it. Because we commission most of the articles that we publish, we have room for only a few of the hundreds of manuscripts--many of them very good--that are submitted each year. We are always pleased to review new submissions, however, and we thank you for thinking of us."

All's well that ends!  Stephen King got over 60 rejection letters before he found anyone crazy enough to publish his stuff.
pyesetz: (flag-over-sunrise)
I wrote:
Some years ago Dr. (X) told me that funding for "programmer's helpers" like (my thesis) was hard to come by.  Funding for "partial Halting solvers" like (my thesis) has always been hard to come by.  Do you think I could get anywhere in the grant department with this project?  I'm trying to plan ahead for when/as/if my Canadian citizenship issues get resolved (perhaps as soon as the middle of next year).


Reply from Dr. C:
Personally I think the work you are doing is exciting, and I think it could be publishable after some experimentation with it to evaluate how it does on real programs.  I'm not sure why he thought it would not draw funding.  If you make it to Canada I'd be pleased to work out a way we could collaborate in this direction. You're doing neat work!
pyesetz: (flag-over-sunrise)
My application has been sent to the Canadian Consulate.  They're a little backed up right now; supposedly it will be about four months before they send an acknowledgment of receipt.  Then it's 2–16 more months until a decision.  Maybe my passport will get stamped by next summer.

I really, really don't want to move to Canada and *then* look for a job, but the only jobs that can be obtained a year in advance are academic.  I've been trying to soften up the chairman of a certain Comp. Sci. department by writing analysis programs for a language he invented.  Here is my latest result.  Note the inscrutable underlying syntax that shines through all the pretty colors and boxes I've added.  His response to my previous emails has been faintly positive; I'm hoping for a "Wow!" for this work but I suppose I may have to be satisfied with an "interesting...".

I've spent the last month on this (in case any of you were wondering why my journal went silent).  The project ran into many problems—I'd be quite surprised to find out that anyone has ever suceeded with this approach before.  The very last problem I ran into today was that my result files are in XML, which is not an acceptable language for pages served from Comcast home-user accounts.  If you put an XML file on a Comcast homepage and then read it with your browser, Comcast's computer tells your browser that the file is plain text, so it looks like colorless gibberish.  Furtopia gets it right, but I didn't want to send him a URL that mentions "pyesetz" because (as explained by [livejournal.com profile] bitch_phd) it's not a good idea to tell a prospective academic employer that you have a blog, even an innocuous nonfurry one.  So I looked around the web for a free hosting service for low-volume non-furry no-advertising web pages.  I almost picked www.cjb.cc, but their FAQ lists specific file types allowed and XML isn't on the list (must be a Windows-based site...).  So I went with freewebs.com.  Like [livejournal.com profile] porsupah, I tend to use the same name at every website, so those of you who know my RL name can easily guess what name my files are under at freewebs.  As for the rest of you, don't worry about it—this crap is all academic anyway.
pyesetz: (mr_peabody)
(Forewarning: in this post I never actually say what my thesis *is*.)

A long time ago, when I was in grad school, one of my tasks was to find a thesis topic so I could write a dissertation.  Topics must revolve around a kernel of a new idea, but the main requirement is that you have to be interested enough in your thesis to keep plugging away at it during the soul-withering anomie of the ABD period.

So I thought up a thesis and discussed it with some potential mentors.  They told me that I should wrap up my idea in aluminum foil, store the foil in a paper bag stapled shut, seal the bag in a plastic sack tied in a knot, and place the sack in a commercial dumpster far from my usual haunts, in the middle of the night, making absolutely sure that I was unseen and thus no one could possibly connect me with this thesis.  Okay, I exaggerate slightly, but I was told (in a quiet voice, so no one could overhear that we were even discussing such a thing) that people who worked on theses like mine either went crazy, or committed suicide, or ran away and tried to pretend that such thoughts had never crossed their minds—and only this last group went on to have successful academic careers.

Having no acceptable idea of my own, I was assigned a thesis topic that had government-grant money behind it.  It was an easy life, but the assigned topic was barren and I didn't think I could ever make a dissertation out of it, so I bailed out with a Master's and left academia.  Meanwhile, the foil-wrapped thesis just sat there, in a dumpster of the mind.


* * * * * * * * * *

In the 1960's, Frank Rosenblatt had staked his career on perceptrons, cybernetic models that he claimed could learn *anything*.  In 1968 Marvin Minsky published a book of proofs showing that lots of things *couldn't* be learned by perceptrons, including the dinky little "XOR problem" that anyone could solve with just three transistors, but wired together in a way that perceptrons couldn't be wired.  That book cast a pall over the entire field of cybernetics: "25 years with no useful results and Minsky says there never will be any".  Research funds dried up.  Rosenblatt committed suicide.  Anyone still working in the area pretended to be doing something else, writing grant proposals for respectable things while secretly continuing to work on cybernetic algorithms.  In 1986 David Rumelhart and James McClelland published a proof that a connectionist cybernetic model could learn the XOR problem.  Rumelhart's paper oversimplified the difficulties and at that time I was cursing him as a poor scientist while trying to replicate his results, but in the end his proof really does (barely) work.  After Rumelhart, researchers came out of the woodwork with results that they had been hiding in their desk drawers because no publisher would touch them.  Today, connectionist methods are commonly used in video games to make the NPC's play better.

Read more... )

Update: Made one attempt to publish this, which failed.

Profile

pyesetz: (Default)
Pyesetz/Песец

December 2024

S M T W T F S
1234567
891011121314
1516171819 2021
22232425262728
293031    

Syndicate

RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Jul. 4th, 2025 02:05 am
Powered by Dreamwidth Studios