No matter how many computers you happen to have, the absolute minimum you need in order to get your work done is at least 50% greater.
I can confirm that this law continues to scale quite effectively over several orders of magnitude. I cannot (publicly) confirm precisely how many.
Reading the Unicode standard for text encoding is surprisingly fascinating. The proposal for the encoding rules for Egyptian Hieroglyphics bring up all the points they’ll need to consider when encoding Mayan Hieroglyphics and various kinds of runes; the proposals for Tengwar and Cirth are just as serious.
OK, I realize this is a strange thing to be doing on a Friday night. But it’s strangely alluring…
A moment ago, after having some remarkably bizarre bit of code explained to me, I walked back to my desk, saying “I am not taking nearly enough drugs to understand this.”
Or at least, I would have said that, but I was interrupted halfway through “understand” by noticing that my screensaver was showing a rendered 3D cow bouncing up and down as on a trampoline, doing interesting flips.
OK, I stand corrected.
From a CERT advisory I received today (CA-2003-22, “Multiple vulnerabilities in Microsoft Internet Explorer”):
VU#548964 – Microsoft Windows BR549.DLL ActiveX control contains
vulnerabilityThe Microsoft Windows BR549.DLL ActiveX control, which provides
support for the Windows Reporting Tool, contains an unknown
vulnerability. The impact of this vulnerability is not known.
Could someone please explain to me what the hell this sort of report is supposed to mean? I mean, was this vulnerability discovered by consulting the Delphic oracles? Or has CERT decided that, in the present legal climate, they can only inform the world of critical bugs by means of gnomic utterances and vague allusions?
I can just see now where this is heading… two years from now, I’ll be getting this —
VU#xxxxxx – Software is all perfectly fine!
There has been a rumor that a certain piece of software has
a minute imperfection. Please do not listen to this at all, nudge
nudge, wink wink. There is no impact to this at all, and you should
not be in any way worried that it could allow an attacker to execute
arbitrary code on ***** systems with the privileges of the root user.
Have a nice day!
For the record: C++ has about half an ass worth of support for actual object-oriented behavior.
I miss Objective-C.
Code runs. Code reads in data. Code runs elaborate curve fits, parametric models, calculates useful parameters. Code then writes the input parameters to the output file, instead of the output data.
Of course, I only notice this after the code runs.
Sigh.
For those of you who had some interest in the OS X clipboard software: It’s now available for download here. No fancy web page yet; I’ll make one of those when I have time. Play with it, use it, LMK if you find any bugs or missing features. I’m especially interested in improving its ability to “guess” data formats correctly – if you find a case where it guesses wrong, or it does something else suboptimal, please let me know.
For those of you who use OS X and the terminal pretty regularly: I’ve got a program that meshes the terminal in with the clipboard (providing copy and paste to files, URL’s and pipes) which is ready for beta. If any of y’all are interested, comment here or drop me an e-mail.
I did some OS X coding last night. The docs for some of the API’s in OS X include a lot of phrases like “The NSString class is ‘toll-free bridged’ to the Core Foundation Kit’s CFString class…”
Sometime late last night, that morphed in my mind into ‘troll-free bridged.’ This led to the thought of an NSTroll class, an NSTrollBridge, and so on.
I’ll just say that too much coding leads to very weird dreams.
Just finished the latest rev of the terminal clipboard utility for OS X. It now can read inputs from files and URL’s, and correctly handle text, raw data, and images. Further formats (PDF, EPS, …) to be added later. It’s starting to be pretty useful, although there’s still a minor annoyance: If your data is anything other than text or raw bits, you have to explicitly tell it so. (If you don’t and copy an image, then paste it elsewhere, it’ll paste the bits of the image as text rather than an image… this has to do with some aspects of the way OS X clipboards work. Need to play with it more to teach it to identify data types.)
If anyone feels like beta-testing it, let me know. It’s handy if you use the terminal a lot.
Update: And thanks to a random inspiration, it now can try to guess what kind of data something is, although it’s a bit slower that way. OK, this is now really in the realm of work avoidance.
Fortunately, I also got some productive work done today, like writing the outline for my oral defense…
Yonatan Zunger’s Blog 