The Haversian Canal

What is the MIT Blog Survey? I took it. What is it? Here's the results of MIT's 2004 blog survey by Fernanda Viegas. The most recent one was administered by Cameron Marlow, also of the MIT Media Lab. If you completed the survey, did you read the consent form? The information page leading to the consent form? Here are Marlow's preliminary results (only available to those who log in to survey).

The real meat is in Viegas' write-up. Rather than quote sections, I recommend you read it. It's short. I'll publish my approach to blogging shortly.

As you can see I now have Flickr and del.icio.us links on the side. So pictures are taken care of and my recent interesting links should be taken care of as well. Now my posts have to have actual thoughts. I have to think? Oh no.

Geena has Grand Rounds, Volume 2, Number 9 up at code blog.

An interesting introduction to the idea that made flickr and del.icio.us such huge successes. Will marketeers manage to fully tap into this data? Will that undermine culture as we know it? Will society fragment into millions of tiny cells? Would that be a culture in and of itself?

Of course, if you're a marketeer-wannabe, here is a short course in conducting market analysis.

Med Gadget is hosting the 2005 Medical Weblog Awards; he's currently accepting nominees.

Professor Philip Greenspun of MIT has accepted my proposal to build an online learning community for the Tulane School of Medicine, at least he's going to let his students consider it. Now I need to come up with what exactly it is a I want. I sketched the web pages as I think they should look. I would be very greatful for your comments by email or posted directly on the sketches.

I finally cracked the code on comments! If you have a TypeKey identity you can now post comments!

Google for haversian canal. It's not right. It's just not right.

Head Nurse recently posted her responses to the ten NCCAM site ethics questions about her site. She went on to say that, for purposes of confidentiality, she felt obliged to conceal her identity in order to protect the identity of patients she writes about. I'm completely up front about who I am.

Since Google can identify its users, since IP addresses can be traced, and whois queries can be made, and all sorts of creative cross checking is possible on Google, community sites, and the good old phone book, it seems to me that perfect anonymity on the Web is an unsafe assumption for any blog author. It is essentially self-censorship. The internet perceives censorship as damage and routes around it. So I assume that everything I write can be traced back to me and therefore I modify my content from there. This is sufficient for any number of journals that publish weekly case studies and images in medicine. Those authors are all to happy to have their names associated with those articles. I think I should be proud of what I write, and I think it is incumbent on me to ensure the patient's identity and not start with the false assumption that I myself am anonymous.

If somebody could tie Berardinelli's movie reviews to a local listings service like moviefone, that would be really nice. Personally, I stick with Berardinelli's reviews and have found them very close to my tastes. If you prefer the modern excess of choice, check out the Online Film Critics Society. Still not tied to local listings.

This is huge! Copyright-safe searching through Google. It's available on the advanced search page.

In the style of a Fermi problem 'never' can be defined by comparing something's half-life to the life of the universe. Examples (somebody check my math: in the spirit of a Fermi problem I rounded a lot):

If a protein of 100 amino acid residues had to assume every possible conformation at the theoretically fastest rate (the period of a molecular vibration is about 10^-13 seconds) then it would take 10^85 seconds or 10^77 years to have a 50% chance of finding its native, biologically active conformation. (paraphrased from Lehninger's Principles of Biochemistry, 3rd Ed.)

How long would a 1000 monkeys have to type to have a 50% chance of punching out Hamlet? There are about 167,000 characters in unique sequence, 96 characters on the keyboard, and a good typist can type 100 words a minute, or about 8 characters per second. The monkeys will be there (96^167)/8 or about 10^320 seconds.

How long would one have to watch an empty room at standard temperature and pressure for a 50% chance of observing every molecule of air in the room spontaneously aggregrate in one half of the room? Assuming air is one, pure, ideal gas, the average speed of these air molecules at 20 degrees C is 300 m/s; there are 6.022x10^23 molecules in a mole of any gas; 1 mole of gas occupies 22.4 liters. Let the room be 5mx5mx3m: all the particles would appear on the left side of the room once in 10^300000000000000000000000 seconds.

Compare those times to the predicted lifespan of the universe, when the last black hole is expected to disintegrate: about 10^80 seconds. Is this a long time? The age of the universe to date is about 4x10^17 seconds.

An interesting lecture on IT conversations by Microsoft Research sociologist Marc Smith.

Sociology provides insights into web communities, essentially collective action through computing. Keywords: "collective action dilemma theory", "interactive sociology", "social network theory", "social software".

What's the opposite of socializing? Getting work done? Well, a lot of getting work done is socializing, so, perhaps they aren't antonyms.

Online community is out (?) Isn't it nice to not answer the question "Is it really a community?" What about groups? Typically too small. Groups are two to eight, ten, twelve people. This 'communities' are 100,000 people.

Key Authors

• Bob Axelrod, Evolution of Cooperation, says we're all engaged in a game, playing risky transactions with each other over and over. The repitition that lets us be social and be successful being social. There is some math behind this called the prisoner's delimma.

• Elinor Ostrom, Governing the Commons: The Evolution of Institutions for Collective Action. Humans are actually good at collective projects; we're only faced with a prisoner's delimma is we treat each other that way. We, in the US, have collectively lost sight of this. Individualism killed the hive? She says there are 7 or 8 things that can be used to fix this.

• Erving Goffman. How do we "do 'being social". By extension, how do we do being social on the net? How many pictures do I need to contribute to be in good standing?

• Edward Tufte, who says pictures really matter.

• Garrett Hardin, who says that all commons will fail because people will take advantage of each other. Perhaps not, not when the membership is large enough.

• Brian S Butler, The Quality of Online Social Relationships. What do we call these things? Are they online communities? Groups? How about voluntary organizations, like the Shriners. 80 to 90% contribute very little.

• What this sites become are virtual 'Shelling Points', a term borrowed from architecture. (Thomas Shelling). Places on a landscape where it seems obvious to meet other people engaged in certain kinds of behavior. One in a million? There's 768 of you on the internet and you can meet at the shelling point. And when we get together, we might even get something done!

• Paul Resnick, Yphrum's Law. Systems that shouldn't work, but do. Like Ebay. If I send money out, then sometimes something will come back. How do you make sure things come back more often? Reputation.

Back to Ostrom's 7 or 8 things:

• Group boundaries are clearly defined.

• Rules governing the use of collective goods are well matched to local needs and conditions.

• Most individuals affected by these rules and policies can participate in modifying the rules and policies.

• The rights of community members to devise their own practices is respected by external authorities.

• A system for monitoring a member's behavior exists; this monitoring is undertaken by the community members themselves.

• A graduated system of remediation is used.

• Community members have access to low-cost, conflict resolution mechanisms."

So there are two general methods for reputation. Graduated system of sanctions, and histories.

There are two things that happen to people when they look for one of this groups? They find to many, and can't tell one from another. When you are walking a city and it's dinner time, you follow rich information to your choice: the movement of other people, smells, colors, shapes, crowds, music.

Then there's a lot of stuff about usenet. Blah, blah. Here's the database report. The experiment appears to have been stopped gathering data in November 2005. More stuff about Marc's predictions of the future.

You can also read Daniel Steinberg's summary of this talk on the O'Reilly Network, and a summary at Julie's Blog.

Unplug the Wacom Tablet's USB connector. Open the Wacom Tablet Properties and plug the Wacom tablet back in. Don't know

I really like the paper in Edward Tufte's books, Monadnock Dulcet. A number of folks have been collectively trying to find out what he uses and how to get it in small sizes and I finally contacted his executive editor, Carolyn Williams, who informed me Visual Display of Quantitative Information is printed on 80 lb paper, and Envisioning Information and Visual Information are printed on 100 lb stock. I need to contact Monadnock about getting on of those two weights cut in the three sizes I want: monarch, 8.5"x11" and 11"x17".

Update, 20 January 2006. Called Lisa in the marketing department of Monadnock paper. She can put you in touch with a local distributor who can, hopefully, arrange for some 80 lb or 100 lb to be cut down for you. Minimum order is 1100 sheets of 23"x35", which is about 150 pounds of paper. A typical box of paper at Office Depot is about 55 pounds. Each 23"x35" sheet should yield eight pieces of 8.5x11. Her number is (603)588-8646.

Olmsted Kirk, (713)868-1531, is Monadnock's distributor in Houston. They can cut to order. Ten reams of 8.5"x11" or five reams of 11"x17" 80 lb Dulcet is $282.35. The paper, 100 pounds of it, is $265.85 and there is a flat cutting fee of $16.50 per 100 pounds of paper. So it's 28.24 per ream for 10 reams of 8.5"x11" or $56.48 per ream for five reams of 11"x17". The heaviest 11"x17" paper at Staples is a 28 lb Hammermill product which goes for $23.65 per ream. My brother is an artist; he said Olmsted Kirk's price sounds very reasonable.

Minus the obvious robots (about 3000 of them), it appears that five hundred attempts were made to submit a comment to this blog in the last 25 days. However, only two comments have actually been submitted. Either I'm misreading my stats and I'm actually just really unpopular, or I'm being terribly rude by denying a lot of comments. I'm don't know, please help me figure this out. If you are reading this, please try to provide a comment. If it doesn't work, please e-mail me at nolson@tulane.edu.

Thanks,

Niels

Interesting.

Better.

Now this is interesting:

Megite | What's Happening Right Now

This site has generated over 600 visits to my site since last night. It's obviously an aggregator, but what, exactly is it doing that beats Technorati by 100 fold? Who uses this? I've never even heard of it.

I just upgraded to Movable Type 3.2. That was scary. Now I have to set up the scode spamware.

Does anyone know how to make InDesign create PDFs with a horizontal scroll bar that moves pages vertically?

Edward Tufte, my personal hero, is now taking orders for his new book Beautiful Evidence, for delivery in May. This represents the culmination of thirty years of thinking about analytic design, the presentation and consumption of quantitative information. His work is a must read, a visual Strunk and White, for any student or teacher of science, any engineer, anyone who plans to touch think about numbers. Appreciators of fine art and excellent writing will surely enjoy the book as well. Anyone who has actually made a book will find the construction of the books themselves exemplary. Much of the draft work has been published on the forum he maintains, Ask ET. Professor Tufte also teaches a one-day course that is absolutely cut rate for the value.

As an added treat, he is also publishing his mom's new book, Artful Sentences. Professor Virginia Tufte is a distinguished emerita at the University of Southern California. I have only one email from the author and two pages of the book to go on, but it seems to build on Strunk and White, sort of Strunk and White's Chapter 5—Style—expanded to provide a Strunkesque treatment of sentences within larger bodies of work. His father also earned a PhD. Regular Wyeth family, these Tuftes.

72 points = 96 pixels = 13 characters = 1 inch

Microsoft is wrong about the layout rules for their own flagship program, Excel. Their conversions between points, pixels and characters boil down to this:

72 points = 96 pixels = 13 characters = 1 inch

13 characters.... what? 13 capital letters in 10 point Arial fit nicely in a 96-pixel-wide column. If the column width is in characters, then one would think Excel would change the character count with font size, but it doesn't. Well, it will if one changes the default font size (Tools/Options, yada yada, and restart Excel) but then a 22 point default font gives a default column width of 8.47 (136 pixels), and the apparant cell width accomodates fewer than 7 characters and just about 4 em dashes. I still don't understand how the character count is calculated, however, this seems quite solid for height and width:

At 100% print scale 72 points = 96 pixels = 1 inch

Of course, the down side of that is one can't draw a rule thinner than .0138" without zooming the print scale and thereby suffering serious typographic layout inconveniences.

Just before starting medical school I did a design project for the National Investment Center for the Seniors Housing & Care Industries. They produce 30 reports every quarter, one for each of the top 30 metro statistical areas (MSAs) in the country. It's 27 pages of statistics in 10 point font. My job was to come up with a design that would allow them to plug an enormous amount of data into a consistent, eye-pleasing format. I was particularly proud of putting upper quartile, mean, median, and lower quartile in descending order, which you can see on page 3 because they're still using my design! Go on, check it out.

I think I'm going to teach myself XML by learning how to contribute to the Google Co-op this summer.

1) Show comparisons
2) Show causality
3) Show more than 1 or 2 variables
4) Integrate word and image
5) Document everything and tell people about it
6) Presentation stands or falls on its quality, integrity and relevance

The deep issue is that PDF is never the editable source file for any document. Portable Document Format is a portable *output* format. It is a core part of Adobe's business to make sure PDFs can be viewed in any operating system and printed on any printer. That PDF is an output format is also why PDF is so popular for distributing documents over the Internet when authors want to preserve their content: the odds of a PDF being compromised without intent are nil.

You can "mark-up" PDFs up and flatten them so they look like clean edits to the casual observer, but you need a full version of Adobe Acrobat (not just the reader), Creative Suite (CS2 is available on education discount for $350, less than any one component retail), Photoshop, or Illustrator. There are some free Linux programs that you can hack together to edit PDFs, but I'm not familiar with them. Google show's a promising PDF editor that actually allows you to edit the PDF's code.

As for converting PDF to DOC, google knows about lots of companies selling software to do that. There may be some free ones.

My upload tools, Mozilla 1.7x and WS_FTP, break off my uploads when I upload files from my home Windows XP box. Other computers, a Dell laptop and a SUSE Linux box, both inside the same router as the Windows box, upload files successfully.

The Mozilla 1.7x interface is barebones, but WS_FTP does some weird stuff. The uploads will purport to start very fast, like 3000 kb/s. The speed then dwindles very quickly. The progress bar very quickly goes to 100% complete but then the status line reads "Timeout, no new data for 900 seconds", after, like, 1.5 seconds, and then the status bar counts down my 5 second retry time and trys again, asking if I'd like to overwrite the file I just supposedly transferred, but don't see in the remote window. If I select overwrite or resume, it will go through this inane process again and again, and again. If I say skip, then it will show me the partial file that was uploaded, which is usually about 350 KB. Sometimes its 570 KB, sometimes, if I try a really small HTML file, like 19 KB, it will be 19 KB, but even then the file won't run, it's partial.

All this only happens on this one Windows XP SP2 machine. This has gone on for three days now. The host server is up and I have uploaded to it with other computers and with other computers inside my home router. The browsers on the Windows box work. I can upload attachments into Gmail. I can upload with the Flickr Uploader. Interestingly, while the Flickr Uploadr takes the usual amount of time to actually post a new photo to Flickr, its progress bar also goes very quickly, very much like WS_FTP's. I have searched this board and Google for various things and tried some of them. I checked that my McAfee firewall is giving full access to all the programs. I've restarted the computer. The Windows Internet Options is set to allow passive FTP. The clock is synchronized to NIST. I reset the Windows firewall to default settings and turned it back off. I turned off the McAfee firewall at the same time. I turned off the router's firewall. None of these usually cause any interference. After all this, with all firewalls down, the upload programs still exhibit the same broken upload behavior.

WinSCP successfully transfers files from the Windows box to the Linux box, both inside the router.

Here's a relevant log excerpt:
------------------------
200 TYPE is now 8-bit binary

[7,8/20/2006 9:56:51 PM,0][2006.08.20 21:56:51.171] PASV

[7,8/20/2006 9:56:51 PM,0][2006.08.20 21:56:51.265] 227 Entering Passive Mode (Server IP address omitted)

[7,8/20/2006 9:56:51 PM,0][2006.08.20 21:56:51.265] connecting data channel to ***Server data channel omitted***

[7,8/20/2006 9:56:51 PM,0][2006.08.20 21:56:51.281] data channel connected to ***Server data channel omitted***

[7,8/20/2006 9:56:51 PM,0][2006.08.20 21:56:51.281] STOR Eye1.wma

[7,8/20/2006 9:56:51 PM,0][2006.08.20 21:56:51.437] 150 Accepted data connection

[7,8/20/2006 9:56:51 PM,1][2006.08.20 21:56:51.437] Send Buffer size: 65535

[7,8/20/2006 9:56:57 PM,0][2006.08.20 21:56:57.359] 421 Timeout (no new data for 900 seconds)

[7,8/20/2006 9:56:57 PM,0][2006.08.20 21:56:57.359] # transferred 7842288 bytes in 5.922 seconds, 10346.026 Kbps ( 1293.253 Kbps), transfer failed.
------------------------
Any suggestions?

Lego Mindstorms NXT is the coolest lego set I've ever seen. Think about it: your budding child engineer can actually use servos and various sensors and the whole thing can walk around! Yeah, it's absurdly expensive, $250, but probably competitve with actually building the thing yourself (yeah, the parts might be cheaper, but you'd have to buy tools, books, etc), and, from a conceptual standpoint, this removes all the bull they'll learn in college and the 12, 13, 14-year old can actually see the product moving around, following commands. The visceral learning value of that at an early age, oh, what I would have given for this...

You heard it here first. From the Web 2.0 Validator:

The score for http://nielsolson.us/ is 8 out of 52

• Uses inline AJAX ?  No
• Is in public beta?  No
• Uses python?  No
• Is Shadows-aware ?  No
• Uses the prefix "meta" or "micro"?  No
• Mentions startup ?  No
• Rocks out to the dance noise of Chinese Forehead ?  No
• Uses tags ?  No
• Mentions Less is More ?  No
• Appears to be non-empty ?  No
• Refers to mash-ups ?  No
• Uses Google Maps API?  No
• Uses Cascading Style Sheets?  Yes!
• Has a Blogline blogroll ?  No
• Appears to be web 3.0 ?  Yes!
• Has favicon ?  No
• Attempts to be XHTML Strict ?  No
• Mentions Dave Legg ?  No
• Mentions an "architecture of participation"?  No
• Appears to use AJAX ?  No
• Refers to the Web 2.0 Validator's ruleset ?  No
• Makes reference to Technorati ?  No
• Appears to be built using Ruby on Rails ?  No
• Refers to Flickr ?  Yes!
• Mentions Chinese Forehead ?  No
• Refers to VCs ?  No
• Mentions Nitro ?  No
• Mentions The Long Tail ?  No
• Possibly contains bytes ?  Yes!
• Mentions Cool Words ?  No
• Links Slashdot and Digg ?  No
• Mentions Ruby?  No
• Has prototype.js ?  No
• Creative Commons license ?  Yes!
• Mentions Ruby ?  No
• Appears to use MonoRail ?  No
• Refers to podcasting ?  No
• Uses microformats ?  No
• Actually mentions Web 2.0 ?  No
• Mentions RDF and the Semantic Web?  No
• Refers to Rocketboom ?  No
• Refers to web2.0validator ?  No
• Refers to del.icio.us ?  Yes!
• Use Catalyst ?  No
• Uses Semantic Markup?  Yes!
• References isometric.sixsided.org?  No
• Appears to over-punctuate ?  Yes!
• References Firefox?  No
• Validates as XHTML 1.1 ?  No
• Mentions 30 Second Rule and Web 2.0 ?  No
• Appears to have Adsense ?  No
• Uses the "blink" tag?  No

Well, a quick Google search (so Web 1.0) shows people are looking for better mobile computing, more use of scalable vector graphics, the overthrow of the Microsoft OS with open source operating systems (any of the various Linux distributions, like SUSE, Debian, Ubuntu, etc) or browser-based operating systems, more semantic design (using ontological languages . . . good luck with that :p). AJAX is pretty much the definition of Web 2.0. If 2.0 was javascript doing work on the client side, then maybe 3.0 will be Flash doing work on the client side, but, as the AJAX folks will tell you, wireframing a GUI-independent design can be painful. Flash seems even more challenged, though you would get obvious design gains among compliant clients. This will be great for the video freaks, the people who want to watch ESPN on their shirt cuff, and "desktop links" (purposeful browsers like Google Earth), but me, I mainly want text, maybe some particularly illuminating illustrations and graphs. Ideas. I thought that was what the Internet was about. So Web 0.0.

[link text](http://linkurl/)

Science magazine recently published the winners of its 2006 Visualization Challenge. Here's my review:

Richard Palais's application of an artisan's craft to pure math is deeply satisfying and his online gallery has many more visualizations.

Nils Sparwasser, who led the Hawaii mountain team, has an online gallery and a new book, Mountains from Space, which Amazon has on discount. One must wonder if he has Imhof's books.

The credit for the radiology images really should have gone to the open source community that has done such an incredible job with the DICOM radiology standard: OsiriX, Dicomworks, etc. I would have to see technical details for the images, but my sense is these particular imagers got their awards more because of their access to particular subjects than any profound ability to design with imaging software, although they may be quite remarkable people in their own rights. By comparison, will Science magazine start giving visualization awards for Photoshopping scientific images?

The Flavio Fenton's cardiac biolectricity tutorials are hardcore modeling and it opened my eyes to what a torsade de pointes arrhythmia really is.

Caryn Babaian's anatomy chalkboard drawing is wonderful, but I would recommend against medical students focusing on artistic accuracy in their drawings. Anatomy classes seem to have a rich, intense first day experience as part of their culture. The students, however, should first worry about spatial organization, nodal analysis, mechanics, embryologic development, and rapid recognition of structures. Drawing a humerus to scale is not the best use of time. It's better than nothing, but make sure you have the logical schema down first.

I particularly like David Yager's green bug because I described the same technique (and surely many others did before I) in the first post I made on this forum.

I can't think of anyone involved in air control who would need Aaron's airspace maps, though I can see why some might want them for their mesmerizing patterns.

The Walter and Eliza Hall Institute has a strong history of supporting informative structural and cellular biology animations, but I have learned more from Xvivo.

I haven't gone through the ornithology site yet, but here's a direct link to their 'browser'.

Here's the Mona Lisa project's flash page.

The only thing I can find on Curtis Dubois is that he may have designed this CD cover.

Didn't find a page for the material informatics group.

Overall, the competition seems to have been more about confections that reasoning about in the real, work-a-day world, and the awards to the radiologists over the programmers illustrates that.

I finally got a laser printer., network printing, and networked filesharing My brother started the process when he came to visit over Thanksgiving. In retrospect, he solved more problems than necessary: he set up my old Lexmark z615 bubblejet in Linux and set up the basic Samba server. He couldn't get the printer working by the time he left, but he got it working by working from the command line, back at his house in San Antonio. Since then I've gotten the Brother 2070N laser printer ($92) he recommended, and that was a relative cake walk. The real savings on the printer is in the ink: the toner cartridges can be had for under $30, and they last for reams, where as we all have our horror stories about bubblejets with $50 cartridges with a thimble-full of ink. And those refill kits? Why should I bother with that when I can have laser-printer crispness for less money? I've also set up all the laptops to access the printer and the share drive, so now there's really no reason to have multiple copies of anything, except perhaps working copies checked out for work on the road where there may be no internet access. And my wife's work, which isn't set up yet. Gotta work on that next, maybe over Christmas break.

Ah, the old on/off switch. So ubiquitous. So overdesigned and thought about in our day and age. It is presented on our computers as the ultimately obvious thing-to-do, fresh out of the box. "Push here". Even a two-year-old could figure that out, and they do. This is annoying with one computer, but when it's the power button for the uninterruptable power supply for a home network with all sorts of file-sharing going on, it's a major time sink to reconnect the network when it goes down. Solution? Put something between the toddler and the power button. How?

I used the cases from SD memory cards on each of the towers (Windows box, Linux box, and UPS box). Just dremel out an access hole in the back, and attach it to the case. Think about it: these cases are made to protect their contents. They are made to not open unintentionally. Speaking from experience, superglue, Testor's plastic cement, and epoxy all fail to adhere to that tiny translucent case. It's got a tortuous rim so it's even splash-resistant. Do I miss the cases? Not really. One card is always in each camera, and one is usually sitting in the reader on one of the computers, so that's three cases readily available.

This is also a nice example of layered security. There's really not that much more between my brilliant young son and the glowing green button, but it obliterates the interesting design of the button, tends to require more planning, and does require some manual dexterity and fine motor strength. All together, it is substantially less likely he will find it interesting enough, and be able to intentionally get to it, until he's old enough to understand that Mommy and Daddy really don't want him playing around with it, which will also be about the time he can remember the causal relationship between that button and the state of the computer.

While it remains entirely possible that he will shove a fork through the entire assembly tomorrow morning, it deterred him tonight (he didn't even go near the computers), and the probability of a two-year-old crashing the network has gone down quite a bit.

12 January 2007: No incidents to date.

7 March 2007: Still no incidents. This fellow pflammerstma on stumbleupon points out that the power button can be disabled in the operating system. Yes, it most certainly can be disabled, and, in fact, it is. The issue is with the UPS unit, and I haven't figured out how to disable the off function of that power button. It's certainly not an available option from the GUI. It was faster to put a hunk of plastic in front.

Can the paper metaphor that already exists on the desktop be extended to the book and from there to the shelf by making your desktop background an image of empty shelves? Would this improve your recollection of where things are on your desktop? You could certainly lump categories of things. Would this allow you to overlap icons on shelves, since, if you know they're all in the same category, maybe you don't need to see the entire icon with it's full name?

The challenge is finding pictures of shelves that aren't at rakish angles, have sufficient texture and detail at multiple levels, and sufficient shelf space. Other possibilities might include paned windows, architectural plan drawings, and small scale botanical garden scenes, but these fail to advance the metaphor. And everyone seems to like blue.

Apple is saying on its site that with Vista out, the next time you get a computer, you might as well get a Mac. Well, I suggest that you should get an empty laptop and get Ubuntu, SuSE, or another Linux of your choice. You can skin it like Mac OS X (Ubuntu is fairly similar to begin with) and, if you're going to use a UN*X-like system, you might as well use one that's free and independent of the platform (OS X originally only ran on PowerPC processors, which are pretty much made exclusively for Apple, but OS X now runs on Intel and PowerPC processors (hanks go to Professor Merz for correcting me).

Microsoft is weak and broken beyond repair.

Apple is nice.

Linux works; it's free, it's powerful, and if you don't like it, throw it out and get another one, for free. Professor Merz, in e-mail, made the very valid point that Adobe Illustrator and Photoshop are not available in Linux and the GIMP is a poor second. I concur with that, but if you don't use Photoshop and Illustrator already, I see no reason not to start with the GIMP and help grow that program through user feedback. If you publish, Photoshop and Illustrator are very, very nice. I will note that, though the interface options are buried, I can use my Wacom tablet in Linux also.

Before I got an UPS, every once in a great while, my desktop would just turn itself off and back on. Much to my frustration. Now that I have an UPS I've learned that every once in a while, the electric grid drops a low voltage in the vicinity of my house. Now my UPS switches over when those happen, and my computer doesn't spontaneously restart. I just hear a couple of beeps when my UPS switches over.

Well, tonight, I was in the middle of doing my taxes when the UPS beeped. Oh, I'd be very unhappy right now if it weren't for that uninterruptable power supply.

I love UPS.

Via Don Hopkins

This is an essay written a while ago (1986 or so) by Richard M Stallman (RMS), about his experiences at the MIT AI Lab, and the story of the Lisp Machine Wars.

Machine Room Folk Dance, Thursday at 8pm.
Come Celebrate the Joy of Programming,
with the World's Most Unbureaucratic Computers.
(There were only five of us dancing, but we had a good time.)

My first experience with computers was with manuals for various languages that I borrowed from counselors at camp. I would write programs on paper just because of the fascination of the concept of programming. I had to strain to think of what the programs should do, because I had nothing to supply me with a goal except that I wanted to program. I wrote programs to add up the cubes of a table of numbers in several assembler languages at various times.