I've never been a fan of ICANN. They're a ludicrously costly organsisation whose "authority" stems from administrative stewardship for some of the internet's universal namespaces. As I used to point out: ICANN costs millions to run every year and their job is to make very occasional changes to small text files - a job I'd happily do for a fraction of the cost.
As someone with some background in net-ops I've always understood that the most elegant bootstrap for a distributed, de-centralized, system is a conservative and tightly controlled hierarchical namespace with a minimally small root. Authority for the DNS namespace isn't granted by governments (with the exception of the US Dept of Commerce) or international treaty. There's no real "root" to the authority ICANN that claims, it's “a consensual hallucination experienced daily by billions” (to use a phrase that others have borrowed from William Gibson).
The authority for management of the DNS namespace is granted by whoever runs the root servers. The root servers are determined by whatever resolvers think they are. The resolvers are usually informed by "root.hints" (a file distributed with Bind) that bootstraps the chain of authority when a nameserver is started.
Who's the kingmaker here? It's just as much the ISC (makers of Bind), as it is the ops staff configuring the nameservers, as it is the company determining their policy of which root to use, as it is the customer complaining he can't send email.
The prevailing attitude is, for the sake of operational sanity, keep using the ICANN root.
Limited experiments in increasing the number of gTLDs (generic top-level domains, as opposed to country-based domains) have been shown to benefit two groups - name registrars and international trademark lawyers - an insignificant number of people as a percentage of Internet users, but whose interests are hugely over-represented at ICANN. I'd argue that, at best, new gTLDs such as .biz have provided no benefit to the public - and are dominated by registrations by either spammers and scammers (hopping from one blacklisted domain to the next) to purely trademark protection registrations. And each new gTLD pulls us further from de-centralization, which (in theory) impacts internet stability.
The theory of "bigger root = less stability" is one I personally subscribe to. Others point to the successful management of ".com", a huge flat namespace, as proof that DNS is robust enough to have a root namespace of millions. They may well be right, but increasing the root of an authoritative namespace like DNS isn't something we can retract. If they're wrong, we're screwed - simple technical clarity that only years of being a grumpy sysadmin can provide.
I've been online on two occasions when the .com servers have failed. My observations at the time were that this did have some knock-on effect to .uk domains - but largely .uk stuff was working fine. Smart hostmasters have authoritative nameservers under at least two different TLDs precisely for this reason. Localization of failure is the benefit of distribution and de-centralization.
With the recent decision by ICANN to flatten the namespace they've shown themselves (in my eyes) to be unsuitable stewards of the root. In my mind (and perhaps others) they're shifting from "de facto root" to merely the "dominant alt-root" - no more legitimate than any other. It's clear that if any authority is going to be claimed, it's going to come from somewhere like the ITU or UN. International bodies like this are, by their nature, slow moving and conservative - exactly what namespace management needs.
And for alt-roots - now is your time. Set up new alt-root servers, grandfather the ICANN root as of this date, then charge hefty administrative fees to include any of the new ICANN-approved domains (registrants can clearly afford it). If you're a large ISP already running a good few resolvers this is a potential new revenue stream - it's practically free money.
In fact I've always suspected it was the secret business plan for OpenDNS.
Reasons I've been told not to take photographs since work moved to the edge of the City of London three weeks ago:
- They were of the back of a police station
- There were children (five to ten metres away, in a nursery just south of Christ Church, which is what I was actually interested in)
- I was standing on the art
- I was on private property
However, in each of these cases, I wasn't asked to delete any photographs (which is, I know, something you can work around, but not if you only have one card). Be grateful for small mercies, I suppose.
When I moved to London eight and a half years ago, it was to start work at a dotcom/consultancy called Oven Digital, based in the old Truman Brewery at Brick Lane. The first time I visited was in late January, and as I walked from the Norwich train out of Liverpool Street there was a clear point where you passed from the City, which clung to Bishopsgate, into a completely different area.
Unlike the glass and steel, it wasn't incredibly well lit, except for Christ Church, spotlit at the end of Brushfield Street. Instead, the low brick buildings on either side concealed a ragtag collection of shops and restaurants, all of which felt like they had a history. The Brewery itself was quiet, off the beaten track, and surprisingly insecure - it was pretty easy to wander around huge parts of the buildings, even when you weren't really meant to. There were vague rumblings about Spitalfields being under threat, but I didn't really bear it much mind. Development happens, right?
Unfortunately, Oven ran into the typical financial difficulties of a dotcom circa 2001, and the London office closed. As I'm not a great follower of fashion, my trips back to the market and Brick Lane have been infrequent and fleeting, so when the employers finally moved into their new office at the beginning of the month, I got to rediscover the area anew, after over seven years.
What have those years wrought? Well, everything is so much more... middle class now. Sure, it's a bohemian middle class, a trendy one, but nonetheless, there's a distinctly different crowd there these days. Of course, where the City's bled east, demolishing half of the old market buildings, that's even more obvious; the business suits give it away. It's also busy; incredibly busy, most days, whereas it used to be that Saturdays were deserted and Sundays the only time it was hard to find space.
More than the people, though, the spaces have changed. As I've said, half the market's gone, although the shell of it around the edges has been retained. The businesses there haven't, though; the family-run pizzerias replaced by SF-based chain makeup stores and expensive Soho bakeries. The worst, though, is that even inside the old market, the character has been almost entirely sterilised away. Bubba's BBQ, which had both supporters and detractors, is gone; Square Pie survived, though. So much for those "free-thinking independents", although I must admit a few (who can, somehow, afford the rent) are holding on. Even the stalwart caff I went to back then, Rosa's, is gone. At least Rossi hangs on, providing a reasonably-priced fry-up.
I think the best single example of this is the Spitz, which moved into the market in 1996, when nobody even heard of it, and which was forced out in 2007 as the final redevelopment of the "saved" market was completed. It's due to be replaced by The Luxe, run by the chef who heads up Smiths of Smithfield. It's hardly going to be the same, is it?
In fact, that's a pretty good epitaph for the entire area. Sure, there's still a Sunday market, and there are retro clothes to be had. The market traders do their best in the scrubbed interior of the retained buildings. To me, though, it'll always be a pale remnant of the place I found when I moved to the city, and the fact it's far from alone - just ask someone who liked the South Bank before the chains moved in - doesn't make it any easier to bear.
At lunch yesterday, I surprised myself by ranting for a minute or two about kettles. Most people don't really think to much about them, but my employers have got through a surprising number of them over the last couple of years, and I've noticed the aspects of the designs that are useable, and those that aren't. Here, then, are a few notes you might find useful if you need to buy an electric kettle.
Upright kettles are far more usable. Sure, retro styling can be cool, but by putting the lid under the handle, they're a pain to fill. They also tend to fall foul of other items below; in particular, the ones we had featured no way at all of figuring out how much water was in them, unless you count "lifting it up and checking it's heavy". Not recommended.
The kettle should be transparent, or have a clear panel. Otherwise you can't tell how much water there is inside. Beware of the "gauge with a plastic ball" design - the ball furs up with limescale (especially in a hard water area like London), and within a few months it'll be useless.
Go cordless with a circular base. Most kettles are cord-free these days - and it is handy for filling - but some still have an asymmetric base, which means you have to fiddle around to put it back. A circular base with a central socket makes it easy to put things back.
Concealed, flat elements don't scale up. Or rather, they do, but at least they're far easier to clean than an exposed element.
In a hard water area like London, you'll need a filter. Most kettles have a gauze filter over the spout, but some - like the Brita kettle candace and I have at home, which is, I believe, discontinued - have a proper water filter. It can be annoying waiting for it to filter, but the water is less full of gunk - especially useful for green teas. (If you're regularly making that, though, you may want to look for a kettle that allows you to set the temperature of the water.)
Obviously, since we've got through a lot of kettles, I have a good idea what the failure modes are. The two common ones are that the power switches break, and that the kettle gets terminally unclean; usually that's because the filter gums up, but it can just be that the elements are too badly scaled for people to be happy. Both are probably only a real issue because of the sheer usage levels of the kettles, but if you're really fond of tea I thought I'd mention them.
For two and a bit years, I had a Canon 350D, the digital SLR that, at times, it seemed every geek owned. I even
wrote a couple of posts here about SLRs in general. Unfortunately, with 35,000 exposures on the clock, it was stolen in November, and I've only just got around to replacing it.
The good thing about the wait is that I was able to skip a generation, getting the new Canon 450D (known as the Digital Rebel XSi in the US), which was announced in January and started shipping late in April. It's been out long enough for DP Review, and presumably others, to review it, and I'm not in the same league, so this is more of an impressionistic review of the first week or so with one.
A large part of the reason I bought the 450D rather than saving quite a lot of money and sticking with the 400D was that it comes with a new kit lens, the 18-55IS (not to be confused with the much faster, much more expensive, 17-55IS). So far I've not noticed much change, but it does seem a little sharper, and the IS is helping a bit with low-light exposures. Certainly it's good enough, for now, that I've barely used the 28-135IS lens that became something of a mainstay on the 350D. We'll see how that pans out in the future, but for now, count me satisified.
Another one of the best features of the camera is one that won't appear on most people's spec lists, namely the viewfinder, and it's gorgeous. My memories of using the 350D were of a dim, cramped window through which to compose, with little in the way of guidance. In contrast, the 450D is much more open, much brighter, and (although this isn't strictly the viewfinder) has more information available to boot - ISO is now shown in the status bar, as well as on the back screen. Along with the lens, this was a key reason for me to choose the 450D, and it hasn't disappointed. Speaking of ISO, there's now an "auto ISO" option even in the "Creative Zone" exposure modes, which means that you can stop worrying about underexposing things so much. Nikon have had this since at least the D70, and I'm glad it's finally percolated through Canon's range.
One more minor change that I'm surprisingly fond of is the option not to rotate images on LCD playback. Both cameras had rotation sensors - a feature I really miss on my Fuji F30 - but the new camera has an option to ignore that on playback, which is fine for me. It's easier to rotate the camera, but see the picture full size, than it is to fiddle with zooming in to try and maximise the viewed area.
What else has changed from the 350D to the 450D? Well, there's sensor cleaning - which was also in the 400D - which I'm really not in a position to judge yet. Mind you, the 350D got quite grubby quite quickly, and anything to help prevent that is welcome. There's also 9 point autofocus, up from 7 on my older camera. Unfortunately, I never got on with Canon's "focus where you're looking" magic autofocus technology, so I rapidly set the 350D to focus on the centre point, and used AF lock to recentre the camera if I needed to. The 450D hasn't really changed anything here, and I was somewhat relieved to see DPReview mention the problem in their conclusions. It's not just me, then.
I suppose the big new headline features are the 12MP sensor and Live View. The former is of little concern to me - most of my images end up on Flickr, and the 350D's 8MP was fine for what prints I bothered with. I was tempted by the latter feature, though, which promises to let you use the LCD rather than the viewfinder for compositions. Well, it works, but the focus is pretty slow. It's probably going to be most useful for astrophotography, where the camera doesn't focus anyway (you have to fiddle with the telescope instead), and that's fine with me. If you were hoping to use the SLR like a point and shoot, though, don't bother. Get a Powershot G9, Ricoh GX100 or Sigma DP1 instead.
Other quibbles? Well, The 450D has a new type of battery, but the charger still just reports "charging" and "charged" - there's no time estimates, unlike the high-end SLRs. It also takes a different type of memory card (SD not CF, since you ask) but since all of those got stolen with my old camera, that wasn't a big deal for me, and since 2GB cards are cheap, you'd have to have a lot of an investment to care. (This is probably a bigger problem for people using the 450D as a second body with a high-end camera than it is to most consumers, who'll only have one or two 1GB CF cards anyway.) I've also found the metering to be a bit problematic, but that might be just me getting used to the camera, and also possibly me misunderstanding the new spot metering feature. I probably need to come back to that.
One final thing that I hadn't realised is that Apple's OS level support for RAW formats doesn't cover the 450D, even with today's release of Mac OS X 10.5.3. There's Canon's own tools, and Camera Raw 1.4.1 knows how to deal with them, but if you're using either Aperture or iPhoto, you're stuck with JPEG for now. No doubt this will be fixed, but it's hard to tell when - it could be in a standalone RAW update soon, or in the next OS revision, which might be three months away. (Edit: turns out that within a couple of hours of posting this, Apple posted RAW Compatibility Update 2.1, so that's that sorted.)
So, would I recommend the 450D? Well, it's tricky. Don't own an SLR, but want one? Now's a great time to buy, and I can't recommend the low-end Nikons because of the lack of a focus motor, so the 450D has to be worth a look. For people with a Canon SLR, it's trickier. There's probably not enough for 400D owners, but for those with 300Ds who haven't yet moved up to the 40D/5D, it's got to be worth considering. If you have a 350D, it's very finely balanced. Personally I'd probably have either waited, or gone with the 40D, taking my old lens with me. However, I can see Canon making just as much money from this camera as they did with its predecessors, and so far, I'm very happy with it.
Today's stupid software idea comes courtesy of the most recent episode of Battlestar Galactica, which featured (as some previous stories have) the Hybrid, the organic controller of a Cylon base star (aka the big pointy bad guy space ships). This week, though, saw probably more Hybrid than any other.
Hybrids, you see, continually babble, a stream of consciousness mixing what sounds like system diagnostics, physics and poetry. After the episode ended, I thought "wait, system diagnostics? Well, if I open up Console, I have those. What if there was an additional process - call it hybridd - which emitted poetry to go along with the more prosaic debugging and whatnot that my computer spits out?"
I have an idea how to do it, too. Algorithm::MarkovChain is a venerable Perl module that puts out almost, but not quite, meaningful sentences, based on an input corpus. Tie that in to the syslog function, a bit of Launch Services, and there you are. (I'm sure you could do a bogstandard Unix version too.)
A further step would be to replace the Console UI with one that boils down the actual computer stuff and tries to fit it in with the hybrid's poetry, but that idea's a lot harder to do well, I'm sure.
Anyway. hybridd - an idea whose time has come. And now, thanks to Tom Insam, here's a Perl version. Requires the aforementioned module, along with File::Tail and Unix::Syslog, both available at your nearest CPAN mirror.
When Adobe first launched Photoshop Express (from hereon in, PX) a month and a half or so ago, it featured integration with three online sites: Facebook, Photobucket, and Picasa. Unfortunately, I don't use any of those for photos. So when I saw that there was an update to add Flickr support, I dug out my old registration - the one that required me to claim I was living in the US, sigh - and had a very quick look.
Flickr appears under the "Other Sites" heading on the left nav - or is it a palette? - of the main window. Clicking the "Flickr" item asks you to authenticate, although somewhat oddly it uses desktop-style auth, so instead of using a nice redirect, PX instead uses a pop up window, which was naturally blocked. It also means you manually have to click about three more buttons than you would with web based auth. Perhaps this is explained by the amount of client-side code, but it still jarred for me. I expect users who don't have to wrangle API auth code would probably cope, though.
Once logged in, Px starts fetching images from Flickr. This is done pretty nicely- pulling each image is slow, so it will carry on if you're not doing anything else, and present the images it's already found for you. Images are sorted by date taken, which is odd if you're used to Flickr's photostream order, which is by date uploaded. (Dates are, naturally, in American format, which annoys me no end, but let's try and ignore that for now.) However, for me it stopped after just over 550 images, which is only about a tenth of the total. I'm not sure why, or how to get it to look for the rest.
Once the images are listed in the Flickr album, they're editable just like any other image available to PX. The tools aren't as sophisticated as those in the main desktop version of Photoshop, Photoshop Elements, or even iPhoto: there's no "levels" tool, and minimal highlight and shadow controls, for example. However, the white balance editor is pretty good, and there's a "Pop Color" effect for those images where you want a red London bus in a monochrome city. Beyond desktop apps, I'd also say that it compares fairly shabbily to Picnik, which is also web-based, but manages a much richer set of tools. Handily, Picnik's integrated into Flickr, making it even more likely to be used.
After editing my image, I wondered where the "save" dialog was, and where I'd get to choose whether the original image was replaced, or who'd get permissions on the new image. It turns out that this is all done automatically. An edited image gets uploaded as a new photo, with your default permissions. The title and description are preserved, but tags and date metadata aren't. To me, this is a killer flaw. Firstly, I want the option to replace an existing image. Secondly, throwing away image metadata is something Photoshop hasn't done since about version 7; it's appalling that PX does this today. Thirdly, I want the option to set privacy levels.
Once again, Picnik's Flickr integration gets all of these things right - in fact, it even seems to have an option to bump images up your photostream with comments intact, which is a very clever trick indeed. In contrast, PX looks like it's hardly trying.
One place that Photoshop Express does try very hard - for publicity - is with images that are copied from its library to Flickr. You can explicitly copy an image into the internal library, create an "album" on Flickr (what's more usually called a set, there), and then copy it back to Flickr in that set. Doing this creates a description that lets everyone know you're using Photoshop Express, and, hey, would you like to use it too? I know everyone is after viral exposure these days, but please let me know you're doing it first and let me set something more sensible.
On the subject of albums, PX loads your Flickr sets as a list of albums, although for some reason this didn't happen the first time I tried it. They're listed in alphabetical order, which, like the "date taken" ordering, is a little odd - Flickr preserves set ordering, and it would be nice if PX would honour that, at least as an option. Opening an album, unfortunately, shows an empty screen, even if there are images in the set. I assume the photo download process is linear. Hopefully a later release will change this, and let the UI take priority, as well as adding caching - each time you open the web app, it has to fetch the list of photos and sets from Flickr afresh.
For all this criticism, I do recognise that Adobe's product is just a beta. On the other hand, given how slick Picnik is, and how nicely it's integrated, it's hard to see how Photoshop Express has much to offer Flickr users, other than a brand name.
Opera just launched their alpha of Dragonfly, "the foundations of Opera's upcoming Developer Tools", which prompted Tom to note that
"Firebug seems to have defined the universe for this lot."
('This lot' includes Safari's Web Inspector, which is actually not that similar to Firebug - the JS debugger is a seperate app for a start - and IE8's developer tools, which I really should look at once I get a disposable Windows image.) It means that all four of the major browsers now ship with developer tools - an impressive change in the last year.
Tom's observation is mirrored by Michael Smith of the W3C: in his XTech presentation when he says that Firebug sets the standard by which all web development tools are to be judged.
The discussion reminded me of Steve Yegge's point in his long, but worthwhile, post about XEmacs in which he said
IDEs are draining users away, but it's not the classic fat-client IDEs that are ultimately going to kill Emacs. It's the browsers. They have all the power of a fat-client platform and all the flexibility of a dynamic system. I said earlier that Firefox wants to be Emacs. It should be obvious that Emacs also wants to be Firefox. Each has what the other lacks, and together they're pretty damn close to the ultimate software package
To be honest, I suspect what he really means here is Firefox + Firebug. At least, if he doesn't mean that, he should be. For me, doing serious web development now requires that combo, even though I dislike Firefox otherwise.
The really interesting point for me is that Firebug, unlike the three other browser development tools, is actually not under the Mozilla Foundation's control. Firefox ships with a DOM Inspector, but this is more of an internal developer tool. Firebug, a third party tool, builds on DOM Inspector's abilities, and it's built for Firefox because Mozilla have developed not just a browser, but a platform. Maybe when I criticised Gecko for their choice to build a platform as well as a browser I was missing something very important.
That extensibility means more than just Firebug, though. If there's a browser that you want to rewire from inside, using the same tools as you do to create web pages (more or less), it's the one from Mozilla. This is where Yegge is coming from, and I suspect that, hidden in a long post that's titled to attract only command-line editor users, it's a point that's likely to be missed.I've never been one of the people who's seen a need to use any terminal application in Mac OS X other than the one supplied by Apple in Utilities. It does the job, uses Monaco 9, doesn't anti-alias (or at least, can be made not to). It starts up in Mac-ish black on white, and generally Just Works.
10.5 saw a bit of an overhaul, with a much saner configuration interface (all in Preferences, rather than hidden in a rather baffling out-of-the-way inspector) and tabbed browsing. Now, I still don't use tabs on the Mac OS, personally; I like the established app/window split and don't see the need to bring a third level of indirection into play, especially when it doesn't even have consistent shortcuts. (Tabs on Windows? Now that's a different story.) In fact, for years I'd quite happily got by with a bunch of scripts in ~/Library/Scripts/Applications/Terminal/ that would neatly stack all the windows.
Sadly, the new version of Terminal also introduced an annoying AppleScript bug which renders these scripts less than useful. When positioning a window, the vertical positions aren't honoured correctly: instead, the window ends up 320 pixels up the screen from the desired location - OK if you want a window at the top, but certainly not if it's meant to be at the bottom right, which is my usual position. I mention this now because the bug in Terminal that broke my window arrangers will also affect a script to centre windows that TALlama (no really) posted in response to a lazytwitter invocation by John Siracusa. If you try to centre a Terminal window, it ends up jammed at the top of the screen, for no apparent reason.
Now, I'm not down with the cool kids who post radr:// URLs, so if anyone who's reading this is, it's really easy to replicate the error: get a Terminal window towards the bottom of the screen, run this script - which should do nothing, as it's merely putting the window back where it started - and watch your window shift around. Do that, report it, and hopefully eventually I'll be able to retire my "set voffset to 320 -- work around AppleScript bug" line.
tell application "Terminal"
set b to bounds of window 1
set bounds of window 1 to b
end tell
Anyway, thanks for listening, and here's hoping for a better Terminal AppleScript interface in 10.5.3.
