All-newer, fast again Weblogs.com
Judging by the quick pings on my last entry, the DNS change for the new weblogs.com server has made it out to me. No more messages about failed pings and read timeouts, and no more trying to explain how a ping could fail when you can see that it worked, because you’re in the update list, right there. Nice.
Fast or not, the comments on my previous entry have got me thinking that I ought to build myself a personal ping proxy server, so that I can have Movable Type just ping it, and then it can take its time about pinging anything I know of that accepts weblogs.com-style pings. Or, for that matter, other styles: a ping proxy server could ping wanderlust.com in a way that they understand, rather than having to send them an XML-RPC message they don’t really comprehend, and then generating an error because they don’t return an XML-RPC response.
I’m trying really hard not to think about how Dave was seeing a heavy load during the changeover because someone was crawling all over the Radio discussion group. Whether he meant radio.userland.com/discuss/ or radiocomments.userland.com/discuss/, in either case if I were crawling it, it would be because of all those juicy email addresses, sitting out at the end of /profiles/$ URLs all over the place. I’ve always thought those were fat enough targets to be well worth writing a special purpose crawler. (Note for the irony impaired: I’m not actually a spammer, or a writer of email harvesters.)
http://www.hiveware.com/enkoder_form.php
http://www.hiveware.com/enkode_tool.php
http://diveintomark.org/archives/2003/02/26/how_to_block_spambots_ban_spybots_and_tell_unwanted_robots_to_go_to_hell.html
I’m torn between loving the tech of Enkode (and wanting to port it to an MT plugin), and the realization that it won’t work either. I’ve been using a slightly less extreme method for quite a while, but that address has still managed to get a few spams, so either someone has written a harvester that can interpret Javascript, or a harvester than recognizes pages where there are hidden (and thus high-value) addresses, and flags them for manual harvest.
I think what I’ll probably do is just take out the email address field in my comments, and the ”why did I put that in there in the first place” email link in the entry byline, link to a hardened contact form, and trust that anyone who comments and wants email contact will leave a link to a page that links to their contact form.
For my site I set MT so that it wouldn’t display email addresses. If the person puts in a URL it will display it, email address just gives the person’s name.
I like it because it seems to give me a slightly added layer of protection in that people seem less likely to leave spam or trolling comments when asked to give an address. I just make sure I’ll never publish it.
I can’t even remember if that’s what I’m doing now, or if I decided that anyone who’s foolish enough to give me an email address but not a URL deserves what they get.
In any case, I can’t agree about requiring an email address to post: in my experience, if you do (or even if you don’t but people don’t bother to look), you just get a ton of comments from people who apparently work for the B Company, since they get their mail at b.com. That has its amusing effects, like the way that (last time I checked) using a at b dot com in any Radiocomments form attaches your name to every comment Mark Pilgrim has ever left on anyone’s Radio weblog, and I think it’s great that you can hardly get a workable email address to ”register” at the New York Times by just pounding on the keyboard, since Mr. alskaweoiupwieqth from apwoeituqpwoirt dot com has usually already signed up, but in practical terms, unless you require a confirmation click on a URL sent to an address, requiring one just requires a fake, and once people are faking you end up looking at the real ones with a jaundiced eye (I failed to reply to someone once, because I assumed that the @hell.com address was a joke, not a real address).
If it works for you, great, but not requiring one has worked so well for me that I think just ripping the field out of the form will work even better.
Here is a ”personal ping proxy” that I hacked together:
http://www.openweblog.info/archive/80614165
Phil,
Now its got that form you wanted as well as the xmlrpc interface. See http://www.openweblog.info/ for more info.
Email Crawlers are Evil
Phil Ringnalda writes about Dave upgrading the weblogs.com machine and finding someone crawling it. I’m trying really hard not to…