It's not just me.
I turned on the Freeview box over the weekend, and pressed the info button. All the programme times for one channel were in GMT+2, and another channel kept showing Â£ signs.
Makes me feel better, somehow.
I am not a Polish pop group...
...but try telling MSN Search users that. Especally the ones that clearly don't understand English, and expect me to understand Polish when they post comments here.
Seriously, it's quite scary. I'm going to defer to another site to actually mention the name, because quite frankly I don't want to make it worse here. Got to 17 in the German top 40 apparently. Don't know about anyone else, but I couldn't stand having the noisy Flash site open for more than about 10 seconds. Meh.
For the record...
Being told you don't have cancer is totally better than being told you do.
Stuff that's irritating me at the moment
- It doesn't fetch robots.txt, and therefore has a hard time following it. I finally managed to stop it indexing my WordPress testbed by blocking it manually.
- So all's fine, right? Well, not really. It's stopped picking up new posts properly. To test, I added myself to my Technorati Favorites [sic]. It turns out I updated 13 hours ago, but haven't made any new posts for 52 days.
- Oh, and the title of the top post on that page was parsed wrong. I think it picked that up from my RSS2 feed, see below.
- Not everything with a feed is a blog. That is all.
Like it or not, robots.txt is a web standard. Well-supported and pretty unambiguous (at least, the "Disallow: /" form is). It doesn't matter whether you're parsing HTML or RSS: If you're indexing data and using it in a search engine, you should follow robots.txt.
Overzealous web standards advocates
XHTML pages are better than HTML pages. If you validate your pages they'll work. Firefox supports web standards perfectly.
I'm not going to go over all the old arguments here. I don't feel I really need to prove that the statements in above paragraph are false. It's not even a complete list, those are just the three that annoy me most.
- I still use XHTML in one place on this site - the Atom feed. It's helpful here, because it means that somebody parsing my feed only needs one parser, and, if they do it properly, is at less risk of screwing up things like entities. In practice of course they'd need a tagsoup parser for other feeds, but that's not my problem (yet).
- IE doesn't support XHTML. Mozilla supports it poorly. And of course Opera and Mozilla are locked out of my pages completely if I make a stupid mistake, so until I'm perfect, or can hire someone to check every page on the site regularly, I'll stick to HTML.
- That's not to say that HTML4 is supported perfectly (it isn't, and I wouldn't expect it to be), but it's more of a case of knowing where I stand. There's a reasonably clear subset of HTML that's been widely supported for about a decade, and just using that seems like the best strategy.
- See the text below the search form to the right? It's too smal in Mozilla, because its parser thinks it's inside the form. This isn't a "Firefox sucks" point, just a reminder that there's no guarantee something valid will work in a "standards compliant" browser.
Okay, I know I'm late. I've finally got fed up of RSS2. Specifically, the people who can't read the spec and don't want it clarified.
The blurb on my blog page now describes Atom as the recommended format, RSS3 as the simple format, and RSS2 as the one-to-use-if-you're-insane format. This is to amuse myself more than anything, but I'll probably be putting in some redirects and denying all existence of the RSS feed at some point.
Oh, and on the subject of Firefox extensions, I made one. Well, sort of. Somebody on moznet (can't remember who, sorry) linked to BBSpot's Top 11 Worst Firefox Extensions, and as soon as I saw it I knew I had to make at least one of them. So here it is: RandomTab. If you use it for any longer than ten minutes, you're certified insane.
EffectiveBrand Toolbars - Spyware?
I've tried their software, in fact I even made my own toolbar (the omgwtfzingbar), but you might want to read this (and the bug above) before trying it yourself. There are certain... inconsistencies... that don't fill me with confidence.
First, the privacy page linked to from the download page. It admits that pages you visit will sometimes be sent to their servers, and says:
To ensure your anonymity, there is no unique ID that can distinguish one user from another.
After installing my toolbar, I found I had a Firefox pref called "CT254130.UserID" with the value "UN20060310144233841". It seems to only be transmitted in order to make the chat feature work, which is reasonable but undisclosed.
Our toolbar: ... DOES NOT download anything on its own. You can manually update the software, but our toolbar will not download anything without your consent.
Any changes that you make to your toolbar will automatically be seen in your users' toolbars. In most cases you won't need to tell them to reinstall the toolbar, they'll see your changes right away.
EffectiveBrand have a page on their website that offers a $5000 reward if you can prove the toolbar meets their own definition of spyware. Doesn't seem fair, but I'll try anyway:
- The program does not install any component or support file that is not expressly required for it's function; including registry modifications.
- The program does not execute any other programs without the user informed consent.
- The program does not launch or initate network connections, unless expressly required by its function or authorised by the user; and does not transmit personal information without user informed consent.
- The program removes itself entirely from the system, including the deletion of all support files and active registry keys.
This is the clearest one. After I uninstalled an EffectiveBrand toolbar, I checked my profile folder, and sure enough, it had left behind 9 files. Whilst I don't really care about the 384KB of wasted disk space, and I'm not sure this is an effective measure of spywariness, I do take pleasure in saying this: EffectiveBrand are spyware producers, by their own definition. And they owe me $5000.
Meh, ChatZilla Stuff again
- 0.9.72 was released, and on time, which is nice.
- Three of the things I mentioned in my other post are now fixed.
- I tried to make a ChatZilla XPI that worked equally well with all of Firefox/SeaMonkey/XULRunner, but it didn't work. Anybody else who has a clue, you're welcome to do it for me...
- I have a review? patch on bug 307774 which should make 0.9.73.
- Speaking of 0.9.73, bug 27807 and friends (network/server list editing) are up for that release. Not sure who's going to do the code, but I'm thinking about it and I'll probably dump some stuff on the wiki later.
Just another quick plug - if you use anything on chatzilla.rdmsoft.com, check the updates page. There's a feed you can subscribe to, as a live bookmark, or whatever, to, well, keep updated...
Scans, sleep deprivation, and sudoku
I stayed up until 2am solving a puzzle. Do I fail yet?
There are little inconsistencies in all programs that are annoying, but not annoying enough to warrant complaining about them. Still, if only subconsciously, they detract from the user experience.
Because I'm in a position to actually fix them, I'm going to list my personal ChatZilla nits that I scribbled down last night in about 5 minutes:
- The scope of the options on the View menu isn't clear. Most are global, Header is view-specific, and Timestamps applies to the current view and all child views. Once the timestamp code is fixed so that a global change doesn't rape so much CPU, evrything should probably be global.
- When editing the topic, there's no Paste on the context menu. Adding it is easy, but I'd like to just make it have the standard textbox menu if possible. Bonus points: I'd like to be able to drag-and-drop to it.
- The logfile prefs are confusing. If I go to edit the log file name for a network view, it shows as "$(network).2006-03-01.log". If I edit the pref to "$(network)-$y-$m-$d.txt", then none of the normal substitution happens and that's the exact file name that is used.
- Help menu isn't very helpful. First requirement: stuff to put on it.
- irc.js refers to ChatZilla prefs once. So does dcc.js, but I'm not sure if that matters.
/mode +b and /mode +e produce ugly output with numerics, repeated channel names, and timestamps.
I'd prefer something like Nick banned *!*@22.214.171.124 from #chan at 19:01 on Sunday [Unban] (with, obviously, the [Unban] missing if I'm not halfop or above).
Some networks have massive MOTDs, and ChatZilla puts each line in a seperate element. This is very slow, and I want to see if it can be improved.
If not, how about a pref to hide the motd unless it changed since last time? Is that at all feasible?
- The new menus have mostly grown on me, but I still find myself going into User Commands when I want an operator command. The submenus just get too fiddly. At least for the userlist menu, where's there's space, the popups shouldn't be used.
- Not a nit, just something I'd like to do if I'm allowed: land bits of my xrmake changes so that, even if it isn't on in the official builds, trunk makexpi.sh can produce XPIs suitable for SeaMonkey, Firefox 1-2, and XULRunner 1.8.
Advance warning: I'm probably going to file bugs on most of these soon. Comments
welcome demanded. :P