net stuff

One issue that Prentiss has emphasized in the past is the need for adherence to a controlled vocabulary when categorizing information. I’ve wondered whether categories could be an emergent outcome of smushing a lot of data together. Well, perhaps, but we ain’t there yet. The chaos at reversible is evidence of that (which I contributed to with my earlier experiment…sorry). Obviously the option exists to take advantage of a useful hierarchy of categories, but the obligation does not.

Preparatory to the appearance of RSS feeds that all the cool kids will want to link to, I’ve tooled up another CSS button:

reversibleTry Reversible: It’s confusing but fun

Later: Word from Joshua is that this will not duplicate Mark Pilgrim’s cool hack: it just makes it easy to create a sort of ad-hoc directory that points back to whatever pages you want. I haven’t asked him, but I suspect this is really almost a diversion he put together on the trail of something else.

I was checking over my error logs recently, looking for problem pages that should be redirected, and found repeated attempts to load FormMail. Now, I do have a form-to-mail CGI on my web host, but not FormMail. FormMail is by far the most popular of these widgets. I was briefly puzzled by the request to load it: I’ve never had it installed, never had a link to it, so I couldn’t quite figure how it could be a bad link.

Until I read a comment on a story somewhere else mentioning that FormMail had a security flaw that made it available for hijacking by spammers. The flaw has been corrected, but it stands to reason there are lots of old installs of it floating around, ripe for the picking.

I don’t quite get it, but Joshua Schachter’s latest project is reversible. This shows where you’re coming from to reach a certain page, as well as where other people came from to get there. OK, that’s nice, but it’s been done before. This also has some kind of categorization system that I don’t understand. I think the real power will come when he generates customized RSS feeds that people can include in their own pages (if it doesn’t completely swamp his server). That will make everyone as cool as Mark Pilgrim.

Later: Let’s try an experiment: Adam Rice | Adam Rice | Adam Rice.

Got my first case of comment spam today. The commenter purports to the have the URL “http://www.1heluva.com/cgi-bin/join.cgi?refer=23911” and comes from IP number 81.86.245.111 (which has been banned).

Don’t bother going to that URL. It froze my browser, and damn near gave me a seizure with all its blinky scrolly bits.

James Gleick, the science writer, had an article on spam in Sunday’s NY Times. Like Jon Udell, I found it disappointingly superficial — I could say the same about some of Gleick’s other writing.

Spam will be hard to legislate out of existence, mostly because of the Net’s global nature. I already receive a fair amount of Korean, Chinese, and Japanese spam, and a little in German. Perhaps, sooner or later, all these countries will pass legislation with teeth to crack down on spammers, but right now, even the USA doesn’t have any.

I already use Spam Assassin, and recommend it highly. It catches most of my spam, but about ten pieces slip past every day (I’m not sure how much I get, but I believe it’s about 100 pieces/day). My mail client, Apple’s mail.app, catches about half of the rest. We could do better–especially because I probably have a small number of false-positives, which bugs me a bit.

Here’s what I’d like to see:

Better collaborative filtering: mail.app already allows you to flag a message as spam, which refines an internal spam-filtering algorithm. This could be extended by submitting the offending message to a central database, which would then push out updates on a regular basis. Such software once existed on the Mac, called Spam Blaster. It was effective, but it was put out of business by Sound Blaster for infringing its trademark on anything with “Blaster” in the name. I believe that Spam Assassin also uses a small number of people to feed new spam into the system, but it can’t be as effective as a massively collaborative system.

Pay: I’d be happy to see a system put in place where everyone pays, say, $10 into an escrow account. If you try to e-mail me and you aren’t on my whitelist, one penny is deducted from your account (I don’t particularly care where the money goes — give it to your ISP). After I receive your message, adding you to my whitelist would just require pushing a button. This would effectively end spam, as spammers couldn’t or wouldn’t pay a penny for every piece of mail sent out. For individuals, though, it would take a long time to work down that $10–that’s sending e-mail to 1,000 new people.

The latter scheme would require some pretty basic architectural changes in the way e-mail works. But considering the gyrations ISPs and individuals are going to already, it’d be worth it. A bigger problem is that it is somewhat undemocratic and bureaucratic: it assumes that everyone has $10 to spare, and the creation of an escrow system (though the system could probably be funded on the interest of the escrowed money).

The former scheme would be somewhat less effective, but could get up and running quickly. Services like Spamcop are doing this now, but for a fee, and inconvenient if you don’t want an @spamcop,net address. It would be worth it for companies like AOL, Hotmail, and Yahoo to run the back-end of a service like this as a free service, simply because they are so relentlessly hammered by spam.

The phrase may be hackneyed, but maybe that’s just because it comes in handy so often. I’ve often said that Austin has been spoiled by its own success.

I wonder if the same thing is happening to Blogger. Let’s be clear: I think Blogger is great. It makes it dead simple — and free — for everyone to start a blog, and in my book, that’s a Good Thing.

The problem is, everyone did, and Blogger has become overwhelmed. People find their archives are evaporating, they’re having trouble posting, etc. I made the switch to Moveable Type a while ago, not because of problems, but because (as I half-jokingly say) my life wasn’t complicated enough. Seriously: I wanted to play around with some of the numerous options that MT offers. A little while ago, Jenny encountered these problems, and so I set her up with a blog inside my own installation of MT. More recently, Dori encountered the same kinds of problems, so I helped her set up her own MT blog (I’m getting good at this). I suspect that the same story has been repeated hundreds of times.

Blogger’s excess of success may be Moveable Type’s success as well. I want Blogger to thrive (I also want it to generate RSS feeds for all its users). I don’t know what needs to happen to keep Blogger running smoothly. The fact that it is free is obviously one of its attractions. You can support it by buying a Blogger Pro subscription, or by buying a Pyrad. Perhaps they need to set up a tip-jar for taking voluntary $5 contributions or something.

This entry at Boingboing recaps (using surprisingly similar language) a comment I sent Greg Elin by e-mail recently. Weird. I’m not accusing anyone of plagiarism — I’m just observing how certain ideas seem to precipitate out of the ether when the building-blocks are in place.

Apparently not — though they did just return, evidently. Ben and Mena Trott, authors of Movable Type (and, apparently, cyborgs who don’t need to sleep) are revving MT again, this time to version 2.6. I’m glad to see them releasing early and often, but a bit chagrined that I just installed 2.51 on a new host. Oh well. I’m guessing the upgrade will be painless enough. There appear to be plenty of tasty features in this release, although the one I’m really looking forward to, user-customizable fields, ain’t there yet. I know it’s on the wishlist, though.

I recently wrote about a new site, GeoURL. In the course of corresponding with that site’s instigator, I also wound up making up the little green badge you see in the obligatory badge zone on this page (and which is appearing in many other blogs, now that GeoURL has been slashdotted).

Some random observations:

There are a lot of interesting things that could be done with GeoURL. First thing that occurred to me is this: create a website where anyone can create a page (sort of like blog meets guestbook?). All they have to do is write up a description of a place in physical reality, give its coordinates, and ping GeoURL. Those places would then show up as links in a GeoURL “neighborhood report.” You could have categories like “park,” “restaurant,” “WiFi hotspot,” etc. Obviously there are problems with this. It would be easy to spam it, so either you’d need an administrator, or you’d need some kind of karma-point voting system (which could also be abused). And some kind of robot-thwarting scheme preventing more than one new entry from a given IP every, say, 10 seconds, and perhaps one of those “distorted graphic” reading tests to sign up. But apart from these implementation problems, this could make interesting things possible. If these categories were part of the tagging for each page, and GeoURL indexed those categories, then one could do a GeoURL search just for restaurants around my neighborhood (for example). This would allow you to bypass Citysearch-type sites with distributed/aggregated tools created directly by regular folks. Hmm. I think many of the tools needed for the front-end of this are probably available already — it’s just a matter of putting them together.

It’s an ego-stroke seeing my little badge being used.

I originally patterned the badge after the XML badge you see here, but I created it using straight CSS markup rather than as a graphic. Joshua (the man behind GeoURL) decided to make a graphic file version of the badge available, and it’s interesting to note that although this is less convenient to put on one’s web page, the majority of the sites using either one seem to be using the graphic. I suspect this correlates to how well their browsers render the CSS: “Oh, that’s ugly. I like the graphic better. I’ll use that.” Or possibly they look at the CSS code and think “Okay, I know a little HTML, but I don’t know what all that gobbledygook is. I’m scared and confused. I’ll use the graphic.” The graphic is actually a screenshot of the CSS, and the two are pixel-for-pixel identical on my screen.

GeoURL is another snazzy way to create a linkage between Internet and physical geography. It helps you create a couple meta tags expressing your meatspace coordinates, which you stick in the header of your blog. You then ping it, and it adds you to its list. This could help make some interesting things possible, apart from just figuring out who your real-world neighbors are in the blogosphere.

This also led me to the ACME mapper, which showed me a 1 meter-per-pixel image of my street, shown here.

Well, this is a first for me. This page played a sound clip when I hit it, a guy saying “This column is brought to you by 3M.”

It’s a good thing I have a fast connection, or I would have resented the extra download time. As it is, it was pretty jarring. If this catches on, it is going to create an entirely new form of annoyance on the Internet.

Yecch.

Time to present some more interesting tools for visualizing abstract relationships. I have a feeling tools like this are going to be much more widely used in the future, and much more intuitive. For the time being, a lot of them are way too slow, and somehow too abstruse.

One that is fast and not abstruse — and has a very useful role to fill is They Rule. This lets you explore relationships among movers and shakers. Sort of like a visual Oracle of Bacon, but for the powerful instead of the famous. Very interesting, although not fully fleshed out yet. Requires Flash 5

There are two similar text-corpus mappers, TextArc and Valence (the latter based on Proce55ing). Valence, technically, is more than a text mapper, but that is one of its tricks. For the time being, these two don’t seem to be so much informative as entertaining, but I can see how visual text analysis could be a serious tool in some contexts. Require Java

A confluence of factors have prodded me into spending the whole day nerding about in the Movable Type back-end.

Movable Type recently went through an upgrade. I was reluctant to install the upgrade, since I had made some custom mods to MT, and wasn’t sure how difficult it would be to reproduce these in the new version. But that new version also has some nice bells and whistles that I wanted, like the handy Search field you see here.

Blogger has been having trouble, and that’s where Jenny’s blog lives (or lived, to be precise). I offered to help her move her blog into MT, and after a week of frustration, she accepted this offer. But I figured “I’ll be damned if I import her old blog, just to wind up updating MT at some point in the future.”

So it was time to bite the bullet. I wiped my old installation of MT off the server (keeping the database, thank you) and installed the new one. Went through the process of customizing it again. Imported Jenny’s blog. Rejiggered my templates to work in her blog with her look. Started messing with the new features. Installed John Gruber’s nifty Smarty Pants plugin.

All things considered, the whole process went pretty smoothly, and more quickly than I’d hoped.

Feast your eyes, hungry readers. I’ve blown a few hours that I shouldn’t have redoing this site. The page you see now uses nothing but CSS for layout–no tables, except for the calendar (which is tabular by nature). I still need to re-do some of the static pages, and I may want to rejigger the structure and the appearance a touch, but I’m pretty happy with it. Let me know what you think (if you care).

Although he probably doesn’t know if, I am in debt to Derek Powazek, from whose page I lifted an extremely obscure, but apparently critical modification. Technical details follow:

I had attempted to use XHTML for my page in the past, but was having a hard time getting the date to line up on the left of the main text. There are some positioning tricks you can use to get it to look right, but those depend on knowing the height of the left-hand slab in advance, which is inelegant at the least. In turns out that by switching to HTML 4.01 Transitional, the float: left/float: left trick would work. If anyone can explain why this makes a difference (and if there’s a way to make the trick work in XHTML), I’d love to hear it.

I haven’t seen this yet, but evidently some enterprising scumbags are now inserting spam into people’s blog comments. A detailed discussion of possible techniques to thwart this ensued.

As a side-dish, I discovered a website that will test how well e-mail addresses on web pages have been obscured to foil spambots (this page passed).

I recently started using Net Newswire, an RSS feed reader. What’s that? RSS is a specialized way of presenting information so that it can be digested by machines, rather than people. The little orange XML icon on my site points to an RSS version of my front page. You can think of an RSS reader as a very specialized web browser for presenting specialized files in a streamlined form.

Anyhow, you get an RSS reader, subscribe to “feeds” that interest you, and the reader sucks in that RSS document and presents it as a menu of stories. You can then quickly browse through a lot of article excerpts. It’s changing the way I read stuff online.

There are a bunch of different RSS readers out there, in the form of websites or specialized apps (even for the Newton!), but OS X readers seem disproportionately well represented.

There are also a lot of RSS feeds out there.

Update: The Guardian just published an article on newsfeed readers.

Disturbing Search Requests has introduced lots of people to the sport of analyzing their server logs for strange referrers. Just for grins, I spent a little time analyzing mine today. Following are some of the phrases typed into Google that somehow brought crossroads.net up as a search result, and (more importantly) that the searcher decided to click on.

pajama pocket speed

dildo shoes

masturbation and air stewardess

smell fear?? maybe you have a leaky nose n not able to smell

good towel head bad towel head photo

creampuff adult video

child pantyhose head

makeovers involving piercing

There were also a bunch of requests that were so specific and spot-on it seemed as if the searcher must have already known they’d find that they wanted here.

Many of you have probably heard the now-old chestnut from John Gilmore that “the Net interprets censorship as damage and routes around it.”

I’m wondering if we won’t see something similar happen with Internet Radio. As far as I can tell, if a webcaster moved operations offshore, he would pretty much be exempt from the recent CARP ruling unless the new host nation passed similar legislation. Right now I’m listening to Radio Liechtenstein, which I suspect has been unaffected. Perhaps some of the other webcasters I have enjoyed but have now been silenced can take advantage of this. There’s an opportunity here.

So I’ve been thinking. Spam is evil. So are e-mail viruses like the klez worm. As are, of course, the people that perpetrate both. Now, one of a spammer’s main burdens is getting the spam out–running software to pump it out. It surprises me a little that no spammer has (yet) used e-mail viruses as a means of distribution. It would make them harder to track down, and alleviate their computing burden (putting it on everyone else, of course, which is the spammer’s stock in trade anyhow).

Of course, viruses are illegal and spam isn’t (yet), which might create a disincentive to use that method. But if it starts happening, well, you read about it here first.

File this one under “web pages that suck.”

I accidentally surfed to the website of a company called IMJ and was blown away at the ingeniously awful navigation bar. It’s flash-based and the items move around under your mouse when you try to click something.