Thanks to all the work put into Blogwise, I’m proud to announce it can now pre-empt blog updates, making this the first blog directory with psychic powers. Sweet huh? 🙂

So what is it with GMail anyway?


D. Keith Robinson asks what the fuss is about GMail anyway. Some
interesting points, and some very interesting responses.

Made me think for a bit. Why was I interested in GMail? Why did I try and
get an invite? Well, I’m quite keen to see what Google has to offer beyond
their searching business. There are many occassions where I think ‘I wish
I had a good, fast search engine’. Email is one such place, and the other
is the desktop – if the articles I’ve seen are true (sorry, no references)
Google are/might be working on file searching and getting themselves into
the desktop application market.

Why have I been prompted to think about email searching? I have a lot of
email; sometimes I need to delve into my archives to pick up email; I run
webmail (portability reasons); most webmail clients that just work off
IMAP are crap at searching.

A Google-like interface might create a fast-access database of my email,
ready for queries such as ‘to:myboss@bigcompany.com pay rise’. If Google’s
speed could be ported over to a single program and database, that’d
outstrip my current search by a factor of at least 20.

GMail offers this, with colourful forms to boot, but it’s run on their
server which has its limitations. Here’s why I’m not using GMail all the

* Google could delete my account or bar my entry at any point. That has
obvious implications about my data on their server. Google Adsense forums
are full of people who’ve had their accounts unexpectedly (and often
without clear justification) terminated.

* i can’t easily transfer my existing mail over to GMail. Best thing I can
do is forward my email from here to there.

* The whole ‘privacy’ thing.. (ads, retaining emails after account
termination, etc.) – although this doesn’t bother me as much as it does
quite a lot of people.

* I don’t want to be svenlatham@gmail.com – I am sven@svenlatham.com!

The first point worries me quite a bit. It might also be true of other
webmail services, although – until now – I haven’t been particularly won
over by other mail services because running my own copy of Squirrelmail at
home seems to have many more benefits. Only when a service with superior
functionality has come along does this become an issue.

I really do hope Google offer a version that offers more options to
resolve the above – probably in exchange for money; I think they’ll get
quite a few customers! I also hope Google push themselves towards the
desktop market. There’s only so much of that irritating XP dog I can take.
It’s smily face holds up for only so long when you’re doing searches on
your own hard disk.

Haiku Results

Wow, I suck at deciding winners for competitions. Next time I’m just going to put GMail invites up on eBay or something… anyway. Congratulations to Matt & David, who’ll be receiving their shiny new GMail invites very shortly. Thanks to everybody, and since I’m such a good sport as soon as I get some more invites everybody else who entered will get one too.

Quick note: Hanni’s hosts seem to have disappeared and taken Pinksocks with them, so her site might be down for a while. You’ll just have to get your regular colourful underwear-related football geek posts elsewhere for now I’m afraid.

GMail Haiku competition

So everybody seems to be giving away GMail accounts lately (those who have one can invite a limited number of friends), and since I’m a GMail user with very few friends* I thought I’d give some away.

At the moment I have two invites to dish out, so you need to do something special to attract my attention. Something with skill. Something that shows you really really want that invite.

…I want you to write me a haiku. A haiku about Google and GMail no less. Leave it in the comments field and the best two I receive by Sunday will get my invites. If you don’t know what a haiku is, have a look here. Or, see my awful example:

“Gmail is so great
Now my inbox is so large
To fit more spam in”

You can do better! Leave your masterpieces in the comments. You can enter multiple times, so go knock yourselves out.

*By ‘very few friends’ I mean very few who don’t already have GMail accounts — of course….

Ooh Shiny

New design is up, and I quite like it. More blog-like than the last one, and fits the screen better. I’ve also put breadcrumbs everywhere and installed a new (better?) gallery. The last one was fine – huge and powerful – but that was also its downfall. It was a git to fiddle with and templating was a no-no. This one is much simpler, templating is a breeze and it’s very hackable. I already added EXIF support to it (okay, so Gallery had that already…)

Anyway, thoughts and comments on the new design very welcome.

Web Capture Redux

Hooray for the Internet. I’ve finally managed to get a script together for capturing thumbnails of websites, for Blogwise’s blog info pages.

Last time I posted about this I was working on a Windows solution, and Brian started putting together a Java/IE combo tool (thanks Brian BTW). Both worked well but ideally I’d like a Linux based solution – that saves me leaving another computer on (which isn’t a popular move in this household).

A few weeks ago I spotted this little program. Webthumb is a Perl script that launches Xvfb, a sort-of dummy windows environment, so you can run GUI-based programs within it but not have it display anywhere. It loads up a browser (Mozilla Firefox) within that virtual screen, and takes captures of a website. A little Perl script wrapper of my own, with appropriate resizing and cropping instructions and we’re good to go.

I also made use of a Kiosk addition for Phoenix Firebirdfox to get rid of the toolbars, although it still needs some tweaking to get it working with the latest version of the browser.

There are still a few slight snags. Popups still appear when a plugin is required (e.g. Flash) but they go away when the next website is requested. I still can’t get Firefox to show full-screen either, so the captures are limited to 640×480, which is less than most sites are designed for these days. so I seem to get the top-left of any given website.

These are all minor niggles though, and progress is certainly being made!

The Big Call

I have a conference call coming up this afternoon, where I’ll be discussing, alongside people from (amongst others) AOL, BT and Yahoo!, how content ratings might be introduced to directories such as Blogwise. Should be good, but the prospect of having a conference call with 13 people seems a bit daunting! The last one I did had 3 and required a lot of phone discipline!

Anyway, more soon. My only regret is the crap land line quality here

Vienna here we come…

Plane tickets are booked, hotel booked, conference tickets booked. Come beginning of July I’m going to the Blogtalk conference. First time abroad for me (I’m a hermit usually…) so I’m really looking forward to it!

Apache and PHP

Ooh this is a tricky one. My self-made templating system which is otherwise coming along quite nicely has hit a slight problem. The system works by running all .html files through a PHP script (via the AddType and Handler config instructions). One of the principles of the templating system is that it integrates quite well with the host server, thus templates are set by .htaccess files in Apache. It normally works great.

When you set all .html files to use a particular handler, Apache doesn’t check that the file exists before it invokes the handler script. If that file doesn’t exist, the handler needs to return an error code to the server, and Apache will then kick off its ErrorDocument for a 404. The snag is, I can’t figure out how to get a PHP module to return an error code to the host server. The best I can do at the moment is echo back a custom 404 page from within PHP, which is not very portable at all – it should be using the document defined in the .htaccess or Apache config.

I’m fairly sure something like this is what I want:
if (!$mydoc=@file($reqfile)) {
} else { ….

But exit(404) doesn’t seem to work. The Apache API docs suggested that I should be returning a HTTP status code as the error, and I can’t find any other information about it. Does anybody know what to do in this case?