“Behold WordPress, Destroyer of CPUs”

My problem is, I know what I’ve decided I think of some things (or people), but I never remember exactly what facts brought me to that opinion. So I know I don’t like WordPress, but I can never find justifications when I need them.

Well, nevermind that there is no cache by default — it’s not that hard to fix with plugins, after all — what I find much more interesting is this sentence:

A default WordPress install will query the database twenty times every time you refresh the page.

I can’t quite believe that the number is accurate, but I can’t find any rebuttal in the comments, and I’m sure someone would have protested. And that’s only an example of what I hate about WordPress: the inefficient queries, the multiplicity of includes, the horribly complicated source code where you have to search for hours (and across fifty files) in order to modify a simple functionality. It may all be very fine when you’ve got a dedicated server, and you can activate the caches for MySQL queries and PHP opcode, but it should never have been released into the hands of a general audience as a DIY solution to be installed on any shared hosting solution.

I never liked the original b2 code base; WordPress has proven itself quite worthy of its heritage, in my opinion. By the way, since you’re asking: I do like the Dotclear code much better (although I’ve never looked at 2.0), but my top recommendation remains Textpattern. Elegant, clever, well-done and versatile. And this wasn’t a sponsored message.

Now, if bad-mouthing WordPress and b2 that doesn’t bring traffic to #FF00AA, I don’t know what will.

“Just a Feed That Works Will Do Nicely”

[…] the nutty practice of offering the same content in two (or three!) different RSS/Atom XML formats. I’ve never understood this — why should would-be subscribers to a feed be forced to choose a specific XML format? It’d be like asking web site visitors what flavor of HTML they wanted: “No XHTML 1.0 Strict for me. HTML 4.01 Transitional, please.”

Well, there’s a historic reason — many newsreaders used to have trouble with one format or another, so you had to know what the application you used was able to process, and choose accordingly. But I agree that the practice is obsolete, and confusing now that almost all browsers have an RSS button. (Doesn’t IE7 automatically subscribe to the first available choice?)

Also, you know there’s a bunch of geeks who would like to have a menu to force some blogs into displaying in XHTML 1.0 Strict.

“L’échec de l’iPhone pousse Orange et Apple à renégocier”

Je suis au bord de la crise d’épilepsie — mon cerveau ne peut pas supporter de tels affronts à la logique et au sens commun si tôt dans la journée.

“Il y a une pression forte d’Apple pour passer à un modèle subventionné,” explique un proche d’Orange.

Cette phrase me donne des tics nerveux, principalement parce que, connaissant Apple, je ne doute malheureusement pas trop de sa véracité. Passons sur le fait que c’est Apple qui a insisté pour vendre l’iPhone au prix fort ; ils ont le droit de s’apercevoir de leurs erreurs et changer d’avis. Mais ce qui déclenche mon eczéma purulent, c’est que l’iPhone est subventionné, et il serait temps que les journalistes arrêtent de répéter qu’il ne l’est pas, juste parce que c’est ce que Steve a dit. L’iPhone est subventionné par les royalties que payent les opérateurs à Apple chaque mois ; juste, il s’agit du seul téléphone au monde à être subventionné mais vendu plein tarif au consommateur final. Et c’est Apple qui vient ensuite se plaindre que le téléphone est vendu trop cher ?

Orange est prêt à subventionner mais souhaite bien sûr une contrepartie. “Le reversement qu’Orange fait à Apple pourrait être revu fortement en baisse, voire même supprimé,” explique une source.

C’est pas que j’aie envie de défendre Orange (après tout, je passe mon temps à vanter les qualités de la Xbox 360, alors au point où j’en suis…), mais pour le coup je suis de leur côté.

“Rocketboom creator pulls plug on auction of Twitter account”

From the “I mentioned it once, I kinda have to follow through” department (I need to have a tag for that):

I believed that I would be able to manage the outcome by trying to make a positive outcome for the buyer, for my friends and followers. Even if it wasn’t a good fit, I (believed) I could work with them. But after I heard that they were all just spam marketers, that just kind of killed it for me and I didn’t want to risk that.

I can’t decide whether it was all a publicity stunt (but that would be daringly negative publicity) or he seriously didn’t foresee that the only people interested in buying Twitter followers were spammers. I believe he may just be stupid enough — that’s why I don’t watch Rocketboom (even though the second host was so much more bearable than Congdon).

Blue Jeans Cable Strikes Back

You may know Monster Cable as the makers of cables that… well, let’s say many people consider grossly overpriced. Turns out they’re also what many people might consider patent trolls, sending cease and desist letters to all kinds of small cable makers and resellers for infringing on their connector patents. (You think connectors are standard and there’s nothing much to be patented about them? Well, that’s the point, that’s why many people might consider they are patent trolls.)

So they ended up sending one of those letters to a small company whose boss happens to be a former lawyer, and he replied with a very detailed legal memo that’s been doing the rounds of the blogosphere; and the reason I’m posting it now is that it does deserve the attention it got (and the company deserves the buzz), because it’s a pretty good read.

I love this bit:

I assume that Monster Cable International, Ltd., in Bermuda, listed on these patents, is an IP holding company and that Monster Cable’s principal US entity pays licensing fees to the Bermuda corporation in order to shift income out of the United States and thereby avoid paying United States federal income tax on those portions of its income.

And this one is sad:

My first seven years were spent primarily on the defense side, where I developed an intense frustration with insurance carriers who would settle meritless claims for nuisance value when the better long-term view would have been to fight against vexatious litigation as a matter of principle.

Yeah, that’s what reasonable lawyers like him end up doing: they quit the business and start selling computer cables.

via bol.cnet.com

Handwriting Recognition for iPhone Now Available

Ah, shit. I’ve thought for a while that one of the things the iPhone misses most is licensing Palm’s Graffiti kinda-handwriting recognition or implementing something similar — I’m pretty sure Graffiti has to be the fastest possible input mechanism for smartphones once you master it. And who cares if it’s not quite intuitive and has to be learned? It doesn’t have to exclude the current soft keyboard; it could only be an alternate mode.

So here’s the “Ah, shit” moment: I don’t want to jailbreak my iPhone, but third-parties will never be able to release such a functionality under the SDK’s terms, so it will never be available to factory-state iPhones unless Steve Jobs decides he likes it. And that’s unlikely to happen.

Okay, I just remembered why I dismissed TextMate again the last time I tried it, even though it had come a long way in finally recognizing AppleScript and different charsets. So this time I’m going to write it down for future reference, next time I see a plugin that inspires me to give it another chance: it’s laggy.

I regularly have long lines of PHP in my scripts and, on my iMac G5, editing them or scrolling horizontally is sluggish — not quite unusable, but far too noticeable for comfort. I suppose it’s the syntax highlighting; sure, it’s much more powerful than TextWrangler’s, but if the result has to be Chinese water torture, I think I’ll pass.

Unlike charsets and scripting, this is unlikely to change by intervention of the developer, so I’ll just try TextMate again if I ever get a multiple-core Intel Mac.

RED (pre-)announces a bunch of cool shit

If you’re not a film or photography geek, you may not know that RED is famous for making a digital camera that makes all the aforementioned geeks drool because it’s so powerful and cheap (for how powerful it is, that is).

They’ve only recently started shipping their original camera, and they’re already diversifying on two fronts: the 5K “Epic” camera (5K means that images are 5,000-pixel wide — full HD television is 1,920 pixels) could be expected as the arms race never stops in technology; the 3K “Scarlet” that looks to fit in your pocket, shoot 3,000-pixel-wide images at a steady 120fps (that means you can shoot true slow-motion at up to 5x) on CompactFlash cards and cost around $3,000, on the other hand, is going to give a lot of people some wet dreams.

WebKit introduces CSS gradients

It’s not as backwards-compatible as most -webkit (or -moz) CSS stuff, but still usable. And, more importantly, it’s going to be very useful in web applications for the iPhone. (Which is obviously why they’re implementing it.)

P.S. Ok, the post concludes that it degrades gracefully if you place several background (or whatever) declarations one after another, but I can’t believe there isn’t a version of IE somewhere that screws up and switches everything to white if you do that.

In the meantime, why doesn’t anybody post screenshots? I don’t want to download the nightly.

via daringfireball.net

Cover Flow in Fluid

If you’ve been wondering why you should be using Fluid, which makes site-specific instances of the WebKit browser (i.e., mini-Safari applications), here’s one reason that’s been true since the beginning, and the only reason I was using it: the best way to use Leopard’s Spaces is to assign each application to a space, and Fluid lets you create individual browsers that you can assign to separate spaces. (I may have already blogged that.)

But now there’s another reason, and I’d dub it “Holy shit” but they call it “Cover Flow,” unimaginatively enough:

If you use a Fluid browser to go to Google, Digg, Wikipedia, Facebook, Flickr and a bunch of other sites, you can display a Cover Flow pane (which you can also set to show in a palette or HUD panel) that preloads all the pages linked from your results list, or friends list, or whatever. Which you can then double-click to open either in Fluid or your system browser, depending how you configured Fluid. And the best part is, it’s customizable — well, it’s supposed to be, but the add/remove buttons don’t seem to work in this version. In theory, though, it seems quite simple and clever: give a URL (with wildcards) the filter must apply to, and tell Fluid the CSS path to the relevant links, and it will fetch the URLs. (For instance, for http://*google.com/search*, it’s #res h2.r a, apparently.)

I only wish it was a Safari plugin instead. But Apple doesn’t want Safari to have plugins. (Isn’t there either a browser or a plugin that displays your browsing history or your tabs in Cover Flow? Well, it’s not nearly as cool anyway. And let’s not mention the search engine that displays its result in a Flash carousel; it’s not comparable.)

Oh, and I know, prefetching links is evil. But when it’s for something as cool as Cover Flow, who cares?

via tuaw.com

Lesson for Flickr: don’t mess with your paying users

I think this is an interesting case study. Any user who is invested in using a service feels a certain sense of ownership over that service - but paid users take this to a completely different level. Paying $25 a year (or $5 a year for that matter) changes the psychology of how users regard a service drastically.

It depends on the website, though: Flickr has always been known for the very 2.0-ishness of its communication with users — being human and fun and laid-back and all. Well, there’s a flip side to that coin, and you end up with a community of users who think they own the site.

Flickr invents “long photos”

I kinda like the angle they chose, limiting clips to 90 seconds — after all, if you extrapolate the traditional Flickr usage, people are gonna want to post either 10-second clips from their cellphones, or two-hour art projects. And you can’t realistically offer to host the latter, so let’s set Flickr very clearly apart from YouTube and all the rest.

The part that sucks, though, is that according to Webware (I can’t find confirmation on the Flickr FAQ, but the example videos I viewed seem to confirm it) videos seem to be converted to 12 frames per second. It’s nice to have better image quality than YouTube, but I think framerate limitations are pushing too far the homage to early-19th-century photography. Why not turn the videos black and white and put an animated grain overlay on top of the Flash players while you’re at it? That would reduce bandwidth usage, too.

(Oh, this post remained as a draft for 36 hours for no reason.)

HuddleChat

Looks like a nice collaboration tool, but I can’t login. I end up on an error page in French even though my Gmail account is set to use English, so God knows what’s happening. Not that I collaborate much anyway.

And you may or may not care about John Gruber’s take:

But HuddleChat is just a feature-for-feature clone of 37signals’s Campfire. The layout is the same, the tabs at the top of the screen are the same, the right-side sidebar listing participants and file uploads is the same. It even copies Campfire’s trick of formatting a message as “code” if it contains literal newline characters.

Borrowing ideas is fair game, but copying an entire app is wrong. And it’s creepy, in a Microsoft-of-the-’90s way, when it’s a $150 billion company cloning an app from a 10-person company.

via daringfireball.net

Google App Engine

I’m a little confused: it doesn’t try to compete with Amazon S3 as a remote storage solution, even though that’s the most compelling part of Amazon Web Services; but it doesn’t really compete with Amazon EC2 either, as it doesn’t provide virtual servers but just hosting for Python scripts — and only that. I know it’s fashionable (but not as much as Ruby on Rails for the web 2.0 crowd), I’m sure the infrastructure will be nice and reliable and well designed, and it’s certainly simpler than managing virtual servers in the cloud, but come on, how limiting is that? Yeah, I know there’s a waiting list already, but that doesn’t mean the App Engine will get much real-world usage.

There’s got to be a point where Google’s traditional “let’s start by launching a half-assed ancillary product to see if there’s demand” modus operandi has to fail. Wait, hasn’t it failed with Google Talk already? And with Gmail even (not that the product itself was half-assed, in that case, but they didn’t try to push it hard enough). And lots of other stuff.

Time Machine via AirPort Disk Is Unsupported

If you attempt to use an 802.11n AirPort Extreme Base Station for Time Machine backup, you won’t get any help from Apple’s technical support, something that readers have already told me. I’ve been getting reports that USB-attached drives work erratically with an AirPort Extreme.

One the one hand, asking people to buy a Time Capsule when they already have an AirPort Express is ugly; on the other, if those setups do work erratically, then maybe there actually are technical reasons. Of sorts. Not to mean “it can’t be done,” but rather something like “our engineers would have to spend some time making it work reliably, and we can’t be bothered.”

via daringfireball.net

Cmd-Minus

Convenient: Firefox 3’s page zoom lets you check how your layout handles window sizes twice as large as what fits on your desktop. Particularly interesting for me, as I tend to like ultra-wide header images these days.

I don’t know whether Opera lets you zoom out the same way, but nobody uses Opera; and I don’t know about the new WebKit because I don’t use the nightlies. But, whichever browser allows it, that’s an interesting use of that functionality.