May 2004 | Main | July 2004

June 30, 2004

Vacation, yay

A whole week of no work! Means I'll actually have to do something for once...

As for Gmail notifications, with tons of help from Darin, I managed to get a pure js/Necko xpcom way of screenscraping working, which means no need for evil hacks. Expect a version that does that when I get back next week.

Posted by doron at 12:54 PM | Comments (28)

June 25, 2004

Gmail Notifier Progress

You can install it for FireFox (Seamonkey was busy doing usefull stuff) from here. Note that it seems that 0.9 itself won't work well, and that you need 0.9 branch builds due to a bug.

Its currently a toolbar item (I'll do a statusbar version next week), and you click on it to login (with an nice login window). After you login, clicking on it will open gmail.

I stress that this is a hack - I am loading gmail into a hidden browser, and when gmail refreshes itself, so will the icon.

The number is the amount of new mails in the inbox.

Free gmail invites for better looking icons :)

Also, it seems that it fails if you have never logged into gmail before with the profile - must be a cookie issue or such.

Posted by doron at 3:13 PM | Comments (139)

June 23, 2004

gmail/mozilla integration

Coming soon to browser near you...

Posted by doron at 1:53 PM | Comments (85)

June 18, 2004

Nokia cash boosts Mozilla

CNet reports that Nokia is funding Minimo.

Posted by doron at 5:01 AM

June 17, 2004

1.7, 0.9, 0.7 - Time to Play the Lottery?

Yup, Mozilla 1.7, Firefox 0.9 and Thunderbird 0.7 are all out - get them at

1.7 replaces 1.4 as the new stable branch. Hopefully people will starting testing it and find all the nifty new bugs we can fix for 1.7.1. A lot of last minute changes went in after RC3 - some being my fault :)

Things that could break: tables and frames (cross domain scripting changes).

Posted by doron at 7:54 PM | Comments (42)

June 11, 2004

Of Popups and Bugs

While surfing I noticed something moving on a webpage, but was focusing on another part of the page. I think to my self "bah, moving popup". Turns out it was a bug crawling on my laptop's LCD.


Posted by doron at 9:56 PM | Comments (1)

June 4, 2004

Backwards Compatability, W3C, WhatWG and Business Logic Markup Language.

I've been reading Hixie's blog and the newly launched Web Hypertext Applications Technology Working Group.

I generally agree with Hixie's comments (other than the Java ones :), and have been for sometime complaining about the W3C's newer standards being overly complex and often missing the point. Someone in my group is working on XML events in Mozilla for example, and they actually made <script> tags with an attribute called "declare" and an ID be referenceable as if they were functions for being event handlers, which makes no sense to me.

Backwards compat is important, even if it requires to keep "evil" behavior. Example is Bug 245274, where I had to revert Mozilla 1.7 in quirks mode to behave like 1.4 did to make sure we don't break stuff out there.

The weakness HTML, XHTML and so forth have is that their widget set is limited - and caused DHTML to be invented, and using divs/spans to create complex controls. XUL obviously added a good base set of complex controls.

Adding more widgets to HTML/XHTML for now is usefull, expecially for accessability, but I think the future lies in create more flexible solutions, and XBL is probably key. It seems the Web Applications Markup Language specification is targeted at that, and I plan on nagging Hixie with ideas/suggests/flames :) Dealing with web applications all the time professionally has hopefully given me some ideas.

One last rant, hixie brings up a good point where a lot of the W3C folks are more interested in server-side than client side:

For example in IBM, a lot of server side stuff is done in Java. They implement tons of APIs, and since Eclipse is Java based, using Eclipse-based clients to talk to that server side is more powerfull than pointing a browser to it.

So what I miss is a way for a server to define its "business logic" (buzzword!) in markup and then send it to the client (be it Mozilla, Eclipse or a console) to process.

For example, it could say that its a tree, and its datasource's WSDL lies at location A, the method to use is called GetFoo(), and define what to do when say an item is selected. I guess it could be called BLML (business logic ML). After seperating style from content, how about seperating logic as well? I think it could fit in well with other standards, especially web services (which serves as a generic client-server connection) and XBL/XUL.

Am I smoking goat weed again? Flames? Love letters?

Posted by doron at 7:59 PM | Comments (6)