Archive for December, 2004

I recently set up an OS X machine, installing 10.3 to start with (base install plus BSD subsystem), and then upgrading to 10.3.7 and installing the required security updates. Having a completely clean system, I decided to install Firefox. I tried to do the install from an end user perspective, seeing what kind of experience they could expect. I figured I was in a position to do this, since I know little about OS X. Here is that tale:

I fired up Safari (great browser, clean interface) and head over to GetFirefox.com to download Firefox 1.0. In true end user style, I decided to forgo reading any documentation whatsoever and just hit the large “Free Download” link.

A few minutes later, I had a Firefox 1.0.dmg.gz file on my desktop, which I double clicked on. I was greeted with the following dialog:

There is no default application specified to open the document Firefox 1.0.dmg.gz. [Cancel] [Choose Application...]

I chose the ‘Choose Application…’ option and tried to find something that could reasonably be expected to open .gz files, but I wasn’t able to find anything that matched that description.

Having now reached the point at which most end users give up, I decided to be a more savvy end user, and go and read up on some docs. So, off I go and I end up at the Firefox 1.0 release notes, which have install instructions:

Double click the Firefox 1.0.dmg.dz Disk Image to uncompress and mount it. Your browser may have already done this for you. Double click the Firefox Disk Image to open it in Finder and drag the Firefox application onto your hard disk. Do not double click the icon in the disk image! Be sure to drag the Firefox application out of the disk image and onto your Hard Disk before running it. Drag the icon to your Dock if you want it to appear there.

Wait a second, didn’t I just do this? Safari had no earthly idea what to do with the .dmg.gz, so it didn’t open it, and obviously MacOS X had no idea either. Dead end.

The Camino download page had some instructions that looked potentially useful:

Once you’ve downloaded the .dmg.gz file, drag it onto Stuffit Expander to decompress it. If the disk image doesn’t mount automatically, double-click on the .dmg file to mount it. If that fails, and the file does not look like a disk image file, do a “Show Info” on the file, and, in the “Open with application” category, choose Disk Copy. In Mac OS 10.2, you can use “Open with” from the context menu.

But, my install of OS X didn’t have Stuffit Expander, nor did it have the Disk Copy application. After a little further research, I found out that starting in OS X 10.3, the functionality of Disk Copy was integrated into Disk Utility, but I was unable to get Disk Utility to do anything useful with the .dmg.gz file. Another dead end. This is way more than the amount of work we can expect 99% of end users to do, and I still hadn’t got Firefox installed.

Now, I realise that OS X has a BSD base, and has command line utilities to un-gzip files. But really, do we expect an end user to have to do that? I hope not. This install routine assumes too much about users’ systems and the tools they have at their disposal. This is not good.

There is an easy solution, which is to make Firefox available as a .dmg file rather than a .dmg.gz. We should be doing this anyway, because the standard way of distributing Mac software is via .dmg files. Presenting the user with a familiar install process will result in a much better user experience, with more completed installs. It’s rather unreasonable to expect end users to have to troubleshoot getting the installer disk image mounted.

UPDATE: The relevant bugs are as follows:

  • Bug 169602: [Camino] Use compressed disk image instead of gzip’d dmg for installer
  • Bug 242845: Firefox disk image should use .dmg internal zlib-compression, not .dmg.gz
  • Bug 268074: [Suite] Mozilla DMGs are redundantly compressed with gzip
  • Bug 276571: Thunderbird disk image should use .dmg internal zlib-compression, not .dmg.gz

UPDATE 2: Firefox, Thunderbird and Mozilla are all now distributed in compressed DMG format!

UPDATE 3 (11 May 2005): Camino nightlies are now distributed in compressed DMG format! Also, Mac OS X 10.4 (Tiger) no longer ships with Stuffit Expander installed by default, thus making this change even more important.

In the last few months, comment spam has become a huge issue, recently culminating in many hosts taking their customers’ MT blogs offline. In related news, Six Apart has hired Jay Allen of MT-Blacklist fame to help them combat comment spam.

I’m hoping that with Jay Allen on board, MT will get some real comment spam fighting tools, including the following:

  1. Integration of supported MT-Blacklist into default MT install.
  2. Addition of a toggle for disabling/enabling comments/trackback in the power entry editing interface, so we can do this in batches.

In a monthly web standards newsgram, HP recommends testing webapps in Firefox. They note that after IE6, Mozilla based browsers are the second most popular UAs that hit HP.com. That’s very interesting, since that means we must be beating out even older revisions of IE.

They put it pretty well: “Make sure your Web Section is as smooth and polished in Firefox as it is in IE. Don’t let your customers find your bugs first!”

It’s great to see this kind of enterprise adoption catching on.

The NYT Firefox ad has gone live. I’m not going to cover it in great detail, since that’s been done already by plenty of others. Instead, I offer you a sample from my Inbox:

Woo!

My name is in the New York Times ad as one of the Mozilla Firefox donators. Ali’s name is listed also.
Here’s the sample ad URL:
http://www.mozilla.org/press/nytimes-firefox-final.pdf

Thanks Ali for spreading the word. I’m going to buy the poster for this thing!

When the NYT ad was still accepting signups, I posted a note about it to an email list consisting of quite a few of my friends from high school, and the above email was seen today on the same list.

As it turns out, Firefox is back in the NYT again and getting more attention, this time in an article by Randall Stross. Some choice quotes from there are reproduced below:

Mr. Schare has said that Mozilla’s Firefox must prove it can smoothly move from version 1.0 to 2.0, and has thus far enjoyed “a bit of a free ride.” If I were the spokesman for the software company that included the company’s browser free on every Windows PC, I’d be more careful about using the phrase “free ride.”

[snip]

Mr. Schare of Microsoft does have one suggestion for those who cannot use the latest patches in Service Pack 2: buy a new personal computer. By the same reasoning, the security problems created by a car’s broken door lock could be solved by buying an entirely new automobile. The analogy comes straight from Mr. Schare. “It’s like buying a car,” he said. “If you want to get the latest safety features, you have to buy the latest model.”

In this case, the very latest model is not an ‘01 Internet Explorer, but an ‘04 Firefox.

Looks like Mr. Schare has a penchant for putting his foot in his mouth. :)

While electronic voting is not bad in principle, the recent implementation during the US presidential election was appaling. Here’s why. I’m by no means a luddite, but if we’re going to use technology to solve problems, lets at least use best practices and leave an audit trail.

The terrorists have already won.

That’s right, they instigated a sense of national fear, duped politicians into passing the unnecessary PATRIOT Act, and have managed to keep Americans scared enough that their leaders (Bush, Rumsfeld, ex-AG Ashcroft, et al.) were able to break down existing protections of personal privacy under the more vague than ever notion of national security. That, and the already opaque US Government has found itself in a position where it can deny accountability for its actions, claim that divulging information would be a threat to national security, and call anyone who challenges them unpatriotic. Really, 9/11 was a gift to the intelligence agencies of the US.

John Barlow has himself experienced (alternate link for those who live in places where TypePad is banned) the extent to which intelligence and security agencies feel they can take liberties. During one of his travels, he was subject to an illegal search of his possessions (the TSA is legally authorised only to search for threats to national security, nothing else), was thrown in jail, and was subject to a body cavity search. I recommend you read his story, which is quite disheartening (to say the least).

When federal agencies feel empowered to take these kind of actions, I wonder what kind of privacy Americans expect to have going into the 21st century? Right now, it’s looking like most Americans are willing to settle for little or no privacy, as long as ‘national security’ (whatever this means) is not breached. By the time people realise what they’re giving up it will be too late.

If the terrorists have succeeded in creating an environment in which its become acceptable to erode personal privacy and relax legal safeguards against government interference in personal life, then I contend that they’ve already won.

I’ve been doing some housekeeping on my blog over the last few days, and have made a few changes and additions:

  1. In addition to the faux blog post about Firefox that IE users see, I’ve added browser sniffing code to my Firefox sidebar box, which serves different content according to what browser you use to view it. Currently, it’s capable of differentiating between (and serves different content to) the following:
    • Firefox
    • Gecko-based, non Firefox
    • Internet Explorer
    • Opera / Safari / Konqueror
    • Others

    There’s no real reason for this, other than one of the PHP tutorials for beginners happened to cover user agent sniffing at an early stage, so I thought I’d tinker with it.

  2. I’ve added TypeKey support to comments. Later on I might enable comment moderation for non-Typekey enabled commenters.
  3. I’ve added a class to my blog called aebrahim-mozillaAdvertising. Any form of static Firefox or Thunderbird advertising is enclosed within this class. Feel free to do whatever you want with it using userContent.css (or whatever alternate method your browser supports).
  4. I’ve made my blog’s HTTP headers a little more sensible. I’ve added Last-Modified headers to all pages, and added charset=utf-8 to the Content-Type header. I’ve also added Vary: User-Agent to the blog homepage.

UPDATE: I’ve removed most of the advertising because I think it’s annoying.

You know you spend too much time thinking about Mozilla when you see the name ‘DoGZiLLa’ on IRC (not on moznet) and it takes you five minutes to figure out that it’s a play on the word ‘Godzilla’.

Why should Firefox steal all the thunder? Now that Thunderbird 1.0 has been released, let’s get cracking and make use of the excellent promotional materials and start letting the world know about Thunderbird. Is it just me, or does the ‘Reclaim your inbox’ button look really sweet? I’ve got a button up on my blog. Have you?

If you haven’t already given Thunderbird a spin, try it today. You won’t regret it.

Congrats to Scott MacGregor and David Bienvenu for this excellent release.

I only recently learned about how you can use different style sheets based on the display media. Since I thought the idea was really cool, I thought I’d have a play around with it. The end result is an update to an article on Bohras. It has different style sheets for screen, print and handheld media, and not only applies different styles based on the media type, but even displays different elements depending on whether its being displayed on screen, on a handheld or in print.

While the implementation is pretty simple, I think it’s functional. I thought I’d post a link to it in case anyone out there is wondering how to do this themselves (it’s not that hard).

The major differences to note between the three versions are the following:

  1. Only the screen version displays the sidebar where article source thumbnail images can be found.
  2. The print version omits some text from the footer which links to the site disclaimer and a link to email the webmaster.
  3. The handheld version does not have justified text (otherwise it looks bad on handheld devices with huge gaps between words).
  4. The handheld version contains no hints on how to display text (wrt fonts and sizes). The handheld UA will make the best decision about this.
  5. Because the sidebar has display: none; placed on it in the handheld version, UAs that respect display: none; will not ordinarily download content linked to from that section. So handheld users won’t download the thumbnails, saving them time (and money, if bandwidth is metered).

In other words, I’ve made some assumptions about what content and formatting is appropriate for each media, and used CSS to put that into effect using separate style sheets (see the <head> of the document). The HTML document itself contains only structural markup, and no presentational markup (except for <em> and <strong>). Because of this, the document has a useful parseable structure (see <h[123]> tags), and degrades nicely on browsers that don’t support CSS at all.