Archive for category Meta
Earlier today, my weblog was attacked by some sort of evil referer spamming porn robot! It loaded my front page, many times per second, with the Referer: header set to the URLs of all sorts of porn sites. It caused my referer list to show all these porn sites, instead of the sites that actually link to me. And it was a denial of service attack, to boot.
A little research showed that the hits were coming from some server at xcite.net, a porn hosting company. So “
Deny from 18.104.22.168/24” did the trick quite handily. The robot finally gave up after eating 403 errors for a while. I purged my referer table of the dirty, and now I’m back in business.
I suspect this was an automated attack. Some idiot must have downloaded a list of all active weblogs from someplace and set his script going. The motive was to get the URLs of his sites to appear in the referer lists of regular sites, hoping that some make their referers public, like I do. This may drive traffic, and possibly the coveted PageRank, to his sites. But not through me.
OK, so you’re all probably out there chuckling to yourselves each time I talk about messing around with the MAME source code to get the music and sprites out of SFII. Well, your little joke is over. I found the “Self Test” menu.
I’ll probably do nicer versions of the MP3s in a few days. But, even with the power of “Self Test”, and understanding how the graphics are stored in the ROM, getting at the sprites is difficult. It turns out that to save space the fine folks at Capcom split up all the graphics into 8×8 pixel squares, and then reused them wherever they could. For some of the characters, all the squares for a particular frame are stored together and are easy to grab. But for others, they’re spread all throughout memory wherever they can fit in. Somewhere I suppose there’s a table that says which squares you use to paint each frame of each guy, but I’m not about to go find it.
But one mystery I always wondered about is solved: When Chun-Li jumps, she does a flip and her feet disappear off the top of the screen. I was always curious if they bothered to draw her feet, even though they are never shown.
The style formerly known as “lite-yellow” is now the default. The style formerly known as “default” is now called “bluerobot”. If you select “bluerobot” it “should” remember your choice.
Here’s a list of hobby type stuff I want to do, written out here so it stares me in the face until I do it. (this method has proved effective in the past)
- weblog stuff
- create an alternate style good enough to be the new default (do you like yellow?)
- fix the URLs so they don’t have ? in them
- render some or all pages statically, especially the various feeds and the homepage
- maybe someday wrap up all my changes and release them as b3 since b2 is no longer active? b2.1? b2-1/2?
- RSS stuff (feed on feeds, my own web based RSS client)
- display DC metadata on each item
- mark individual items as read instead of all items
- cron job to update every few hours when I’m not around so I don’t miss any items
- release to unsuspecting world, hear crickets
- recompile MAME to be a sprite and music ripping monster
- the whole world domination thing… I haven’t forgotten
Check it out now! See over on the right where it says “styles”? Try yellow, then switch back to default. Isn’t that cool? It should remember which one you chose and use it for future visits. Thanks to A List Apart for figuring this out for everybody.
(I already know that on Mozilla/Mac, if you switch to yellow and then back, my name turns into Steve Minutill. Other bug reports welcome.)
Here’s my new idea, now that my previous two are done. Not really an idea for this weblog, but a project I want to do.
- It isn’t integrated with my browser well enough. Sometimes I want to read news items right in the newsreader. Sometimes I want to open up individual items in a page. Sometimes I want to click on links in news items and open tabs. Sometimes I want to go to the home page of the site that the news is from since their feed is plain text only. All these things are inconvenient/impossible in Feedreader.
- I don’t want it to sit there reloading feeds all night when I’m not there. There’s no way to “pause” it, all I can do is close it, and then it loses its place the next morning and everything is marked unread. Not cool. This is a minor problem, I know.
- A bigger problem: What about when I get home? If I used an RSS client here too, it wouldn’t be in sync. Stupid.
To get around all these limitaions, I want to make a browser based RSS client. Since it’s right in my browser, all the different ways of navigation I want will just work. I’m thinking of making it only update the feeds, if they’re out of date, when I load the page, so it won’t sit there stupidly updating all night. It will then be in sync wherever I go: I can just load my feeds page and see the latest news, kind of like Livejournal’s friends pages, or News Is Free. But I want it to keep track of what I’ve read, and of course be free and Free. If I do it, I’ll make the code available so others can set up their own versions on their own servers.
I would set it up here so random people can log in and set up personal lists of feeds, but I don’t think minutillo.com could handle the strain. Maybe if it gets good enough some nice person with a big honking server will set up a public version, ready to go.
I think with a little PHP, and a little MySQL, and Magpie to do all the heavy lifting, it shouldn’t be that hard. Wish me luck!
Just the other day, I was going on and on like I usually do, this time about how I don’t care how this site looks in Netscape 4. Well, today one of my friends told me that A) he is stuck at 4.x at his technologically-backwards place of employment and 2) my site like TOTALLY DOESN’T WORK.
This will not stand.
So, I’ve Zeldmanized it! Yeah that’s right, on 4.x browsers you’ll get all the content, with none of the style. It’s like being back in 1994.
Those of you on Mozilla-based browsers can simulate this effect, if you like, by saying View|Use Style|Basic Page Style.
The URI of each item is the exact referer URL I saw.
The <title> is the search query they used.
The <link> is the URL on this weblog they ended up at.
The <description> is the domain name of the search engine they used.
I think this is a cool idea because:
- I spend too much time eye-grepping through my logs seeing what search terms people are using to find me. Now the information is available in real time in a convenient format.
- This is the big one: Imagine if all weblogs, or even if all sites, made this same information available. That would mean somebody could aggregate them all together, and we’d have the equivalent of the Google Zeitgeist, except updated in real time, and the data would be Free.
I had two ideas for this weblog recently:
- Create a RDF/RSS/XML/whatever feed of search terms that led people to this site. I’ll explain later why I think this is cool.
- Create some sort of department where people can see the code running this site, with my mods to the standard b2 distribution highlighted somehow.
I am sharing this with you so that the longer I am lazy and don’t implement these ideas, the more embarrassed I will become.
Memeufacture is a neat idea. John Wehr is scanning weblogs (probably RSS feeds) and seeing who links to what, and more importantly, who links to what first. Then, he’s figuring out which weblogs tend to have the best (most propogated) links earliest. These he deems to be “most influential”. Right now in the popular category, the top three most influential are Blogdex (which is another meta- site and should be filtered out, or maybe this problem will take care of itself as he adds more “real” weblogs to his spider list), Slashdot, and Boing Boing. Sounds about right.
This is another important tool for those of us not satsfied with getting our memes recycled, second or third hand, and who are always backtracking up the river, following the flow, and trying to find the True Source.
Check out the bottom of the sidebar over on the right: I added a new section that shows the 5 most recent comments, with links to the associated stories.
Next up to fix: Make the permalink and archive URLs look nicer. Get rid of question marks. And fix the Last-Modified and ETag headers to be more accurate by incorporating the most recent comment post date as well as the most recent post so people don’t keep double posting all the time.
I think there was either a bug in b2 or maybe in my configuration, but the ETag and Last-Modified headers were all screwed up. Minutillo.com runs on GMT, but I want the dates in my postings to come out in my local time, so I told b2 what the offset was. It seems b2 adjusts the dates before it writes the dates into the DB, which means when it gets the date of the last posting to generate the Last-Modified header, that date is in EST. That’s not allowed, according to the this section of the HTTP RFC. So a little hacking (reverse the offset) and I’ve fixed it, I think. I’m not confident enough about any of this to submit a patch back to Michel, though.
Maybe I’ll let him know in a few days if this fix seems right.
One of the subgoals on my way to world domination is to make sure that this site comes up first when you search for Steve Minutillo on Google. Well, that was easy. What’s more surprising is if you just search for Minutillo, I come up third, right after this Hana Minutillo who is apparently some sort of opera singer. On that same first page:
- A doctor (Angelo Minutillo)
- A jeweller (Laurel Minutillo)
- A professor (Patrick Minutillo (interestingly, there’s a good chance my name would have been Patrick Minutillo had I been born one day earlier (I was born on March 18th)))
- An eerily familiar looking guy who runs Minutillo’s Performing Arts Conservatory (Lenny Minutillo)
- And of course this site.
I wonder if my ranking will be hurt when I fix the problem that makes it look to Google like all my site says is “STEVE MINUTILLO STEVE MINUTILLO STEVE MINUTILLO STEVE MINUTILLO STEVE MINUTILLO STEVE MINUTILLO”? I was doing a stupid CSS trick.