Internet Explorer 7 beta 2

IE7 beta 2 is out, and all the MS bloggers are pimping it, so I thought I’d actually check out the site (I might install it at work, but there’s no chance I’m installing it at home).

So you visit the IE7 homepage and you’re greeted by a terrible mid-1990s flash animation touting IE’s new features, all of which sound quite remarkably similar to features everyone who uses Firefox and Opera has been using for a few years now. Finally, you’ll realise that this isn’t actually a splash page that is going to forward to actual content, but the scrolling words are clickable items and then when you scroll down (my monitor’s currently at 1024×768, I have two large-ish toolbars and the navigation menu for the page isn’t visible in Firefox or IE6 – didn’t we solve scalable Flash about seven years ago?) you’ll find the actual navigation, from where you can get to the tour.

I was actually looking forward to the tour, I ignored the first IE7 beta because it was bound to be terrible, but now I’m keen to see the new cutting-edge features:

Tabbed browsing! Streamlined user interface!

Oh dear.

Instant RSS feed detection! Toolbar search box!

Oh deary dear.

Unless they’ve got something spectacular to pull out of the hat, I think I pity the IE7 developers. I mean, they’ve been working for a good couple of years now, and they’re just approaching Opera and Firefox’s heels? I haven’t seen much of the Internet’s response, but I think they’re going to be ripped apart for this. I initially took Rod Begbie’s comment with a pinch of salt, but it’s dead on:

Microsoft have released a public beta of IE7. The homepage is a staggeringly fucking awful flash animation, promising “Everything you need, nothing you don’t, and a few things you have yet to imagine”, before displaying a list of Firefox’s features.

Running a CMS on Atom

At work I have the need to very quickly design and roll out an insanely basic CMS for a website.

There is no official workflow, all content will be in XHTML, people need to be able to create new documents, and load, edit and save existing documents, with some degree of metadata surrounding it.

I’m currently pretty convinced that my editor will be TinyMCE.

My first thought is to write a very simple Atom backend, where each page on the website is stored as an entry in a feed along with appropriate metadata, and the editor creates new entries or edits existing ones.

The problem then is that pages appear within sections on the site. Sections can be easily defined as categories in each entry, but how do I administer the list of sections itself? At some point when creating or editing an entry, a user will need to be able to choose which section the entry belongs in; this implies to me a separate list of categories, perhaps stored as XOXO ? But then what happens to all the pre-assigned entries if I change the name of an item in the XOXO file?

What about if sections are physical directories, each containing a single Atom file, which contains the pages for that section? Or is that too much overlap between the abstraction and the implementation?

Any hastily-thought ideas and suggestions welcome!

del.icio.us greasemonkey script

A couple of weeks ago, Tom Coates was moaning about a UI aspect of del.icio.us:

I do not understand why del.icio.us’ submission form doesn’t limit the amount of characters that you can input into it. Instead, you’re left to work out later that whatever you’d written had been cut off automatically at a couple of hundred characters to fit into the database. This must be the simplest thing to fix, surely?

Actually, yes, it is. Here’s a Greasemonkey script which limits the description field to 255 characters and displays a character counter after the text box so you can see how many you’ve got left when adding a pithy comment.

Install del.icio.us description

Why would I want to edit in OPML?

I don’t get all Dave Winer’s fuss about OPML.

Now he’s trying to get his OPML Editor tool officially integrated with WordPress (an addition to the ability to create new blog posts using wordPress.root, which incidentally, I can’t get to work) so that you can edit your “About” page in it.

Why would I want to edit my “About” page in an OPML Editor? I mean, seriously? I could understand possibly wanting to edit my categories and blogroll in a native OPML editor, but actual content? I don’t buy it. I know the argument is that each blog post is effectively a tree of paragraphs which contain other elements, but not only is that not the way I (or any other human I’ve met) thinks, but it also seems counter-intuitive to create and edit in that way.

At this point, someone will say “but my mother/grandmother/aunt/little sister can use this! they can’t use XXX!”, which is such rubbish from so many different angles it’s not worth debating (again).

So, quite apart from not actually understanding why on earth you’d possibly want to do something like that (which even the normally sane Robert Sayre agrees with, so I could be missing something) but dear god. The OPML Editor. Have you used that thing? Jesus christ, it’s an abomination of not only usability but design and human cognition. Some guy in Dave’s comments says that it’s open source, but I can’t see that source anywhere on the download page. At least that would give me the opportunity to make it better if I thought it was a worthwhile venture. Which I don’t, by the way.

I think the OPML Editor wins my 2005/06 award for “UI most likely to make me cry”.

Tello extends presence to SIP

Presence has been gradually seeping out from IM over the past few years into other applications, notably in the XMPP/Jabber community where presence is a key (it’s the first ‘P’ in XMPP!). The benefits of using XMPP are mainly that it’s an open, documented and extensible standard which allows anyone to make use of it.

In an article called Say Tello for Presence Om Malik writes about a new VoIP focussed start-up called Tello, based in San Mateo, California. Doug Renert, a former executive at Oracle Corp heads up the start-up..

The company, The Wall Street Journal writes, is going to allow users to “workers see on their computers or mobile devices whether the person they are trying to reach is on an office phone or cellphone or is logged on to instant messaging.” [Read how it works over on Business Week.]

They do this by installing SIP proxy servers so that as well as picking up your IM presence they can detect your telephony presence. This is an interesting innovation, and one which, had I been paying attention to this market, would have been immediately obvious.

What’s more interesting is that this kind of thing could be provided by an XMPP-based application but probably isn’t. On the jdev mailing list we often hear about people implementing custom internal solutions utilising XMPP for backend communications and presence, but it’s much rarer that you’ll hear about large companies deploying commercial solutions based on XMPP or even startups. Of course, this could be for two reasons:

  1. I’m just not hearing about them
  2. Companies are reluctant to talk about the technologies they’re using, so there may be lots of XMPP commercial application but we don’t know about them

The latter is entirely possible, and in fact quite likely. The obvious recent exception is Google, whose Talk application not only uses XMPP but extends it in a well-documented way and provides example interop libraries.

The question though, is are there any others? Could Tello be using XMPP? Should this kind of solution already exist in XMPP so that, for example, I could start using it at home?

Gmail’s new delete button

Gmail has rolled out a delete button, so that users can now easily delete their mail instead of just archiving it (because despite what some people think, you could actually delete your mail before via the drop-down box of actions).

This has come after a pretty long period of complaining by Gmail users, and the addition has been widely heralded as a good one. I have one message to those people: you don’t understand Gmail.

Honestly, if you think you needed a “delete” button (or were using the Greasemonkey extension), you’ve completely misunderstood not only the Gmail interface but also the new experience the team has tried to give you. No exceptions. If you disagree with this, and also think that, like someone in the betanews comments, Now If they will get rid of the Archive thing and star thing for something better it would be even better, you should stop using Gmail for web-based mail. Seriously. Or if you do want to keep using it because of the 2GB space, start using a desktop POP client like Thunderbird. That way you’ll be completely immured from what Google are trying to do, and you can go back to your 1995 interface.

If this isn’t obviously apparent to you, just take a moment to actually read that part of the Gmail help explaining delete. For a start, it’s called “I’m not sure if I should archive or delete” – bingo. Such equivocation is totally unnecessary for all except the most pedantically set-in-their-ways; “archive” is enough. The text of the help even goes on to encourage you to use “archive” rather than “delete”.

I feel so, so sorry for Gmail’s developers.

Kinja’s new look and new features

In December 2004, I wrote that, despite really liking it, and despite it being part of Gawker Media (have they been taken over by News Corp. or not? I can’t tell) “Kinja is the web aggregator everyone forgot“.

Well, in December 2005, Kinja got not only a new look (for the worse, I think – it’s now far less obvious exactly how to traverse the site and how to do things), but also some new features (although the blog doesn’t actually list them, the dumbasses – but then again, there weren’t any posts to it at all between June and December, so what can you expect?).

One of the more interesting new features is that every blog Kinja tracks now has it’s own “Kinja Card”. Mine is on http://www.kinja.com/id.knj?url=philwilson.org/blog, you can change the url parameter to look at whatever blog you like. As well as linking to my last few posts, it includes my Alexa traffic graph, posting frequency, links to my blog from tracking sites like pubsub and technorati, and links to posts from other blogs Kinja judges to be related to mine. You can also export this list of related blogs as OPML which is quite interesting. You can also export the feeds you’re currently subscribed to as OPML.

All in all, although it’s clearly trying to deliver more features, it now feels like Kinja doesn’t know whether it’s an aggregator or a tracking site, or a site for tracking tracking sites. If that makes sense. It’s a shame though, because I really think that they had something going on over there.