A quick update: I got the memory upgrade from 512MB to 2GB for my powerbook, and it has helped performance hugely. Meanwhile, Maya's MacBook had become similarly unusable in much the same ways as mine--a lot of quality beachball time--despite her using it for nothing other than web browsing and word processing the vast majority of the time. So I got her the same upgrade. Fortunately the memory was pretty cheap. But this is a bit ridiculous. I'm not sure who to blame: Apple, Firefox, or people who make websites. Should a MacBook out of the box be insufficient for web browsing?

Two sites that could really benefit from massive Google Maps mashups are TripAdvisor and MenuPages.

Menupages is a pretty great site for finding a restaurant. You can tell a lot from the restaurant's menu (no more having to deal with vague notions of $$ and $$$) and customer reviews (though it does entail some wading through shills and people who seem to dine with the aim of picking a fight). The one real problem I have is that its neighborhood divisions are far too coarse. Most of the time I have a specific place that I'm already going to be in, for a movie or whatever, and I need something within pretty close proximity, say 3 or 4 blocks. I choose "Upper West Side" and I get everything from 59th to 115th street, from CPW to the River. Hundreds of restaurants. What I desperately need is to be able to sort the list by proximity to my given address. And if they could do that, it would probably be trivial to have a Google Map of everything.

Google does have their own local business finder which can be used for restaurants, and it shows you a pretty good smattering of reviews at a glance, but their system has its own annoying quirk. It also refuses to show you the closest restaurants to the address you provide. Instead, the first page of 10 results will have maybe one or two that are quite close, then the rest will be scattered around up to 10 blocks away. Go to the next page and it's a similar distribution. The last time I used it, it seemed to choose the most expensive restaurants first, but I can't say for sure if it always behaves this way. Perhaps they chose this to avoid maps with too many little pointers cluttered together, but I don't think it's the right compromise.

* * *

Tripadvisor has become as indispensable in my travel plans as MenuPages has for my dinner plans. But I end up in the same routine--read about the place, and if I like it, check it out on Google Maps for proximity to public transportation and places of interest. In this case I sometimes have to massage the address to deal with language and local street naming conventions. I'm sure for both of these sites it would be a large and tedious project to check all their addresses and ensure proper placement on the maps. Existing agreements could pose a problem as well; I see that MenuPages at least has Mapquest links. There may also be limitations to how many locations can be placed on a Google Map. But if it's technically feasible I think the benefit to usability would far outweigh the costs.

PS I'm amazed how many sites still used maps from Mapquest, Microsoft Mappoint, and others even when Google maps had blown them all out of the water--did they all get roped into onerous ten year agreements, signed in blood? I admit the others have caught up a bit now, but they're still not as good.

At A List Apart, the article Graceful E-Mail Obfuscation makes me think about a funny aspect of the spam war. There's a neverending march of articles with the latest and greatest method for defeating spam in its many pernicious forms. Individuals have long used their own favorite method of obfuscating their addresses when posting them in public places (although it seems more common today to simply hide them entirely and use safer methods of communication, such as private messages on message boards). Spam and email address harvesting are volume businesses, so it's true that even if your address is easily parsed by a human, a harvester (the person) would never bother with manually searching it out, and they probably won't bother trying to make their harvesting programs smart enough to parse every possible obfuscation. But if a particular method, such as replacing @ with (a), becomes popular enough, it might become worthwhile for the harvester to gain a competitive advantage by enabling their program to parse it. The same is true for many other cases, such as spam filters in email programs.

So in a sense, the best way to ensure the defeat of an anti-spam method is to popularize it. The only way to win the war is for everyone to have their own custom anti-spam measures, such that there's no efficient way for a spammer to outwit all of them. On some level people already do this--they obfuscate their email addresses in their own funny little way, and lots of people who run web sites come up with their own little system or clever CAPTCHA. In other cases, such as email filtering, it's far less practical for individuals to take the time to do this. This is a problem that resists the incredible efficiency of distributing software--people writing programs or plugins or writing articles with some html code are doomed to be defeated by their own success, while anyone creative, self-motivating, and perhaps selfish enough can win.

Today I happened upon Damien Hirst's new work in the lobby of Lever House, across the street from my office today. I later found out that the gala opening was on Saturday night, but I'm glad it caught me by surprise. The effect was pretty dramatic.

Comments overheard at the exhibit:

"Those definitely weren't chirping yesterday" (in reference to two live birds in a cage)

"I'd like to meet this guy in a dark alley and smack him around."

Herald Tribune story here.