Plod

#1. The concept of "language packs" that was so prevalent on the Mozilla 1.x and Netscape 6.x / 7.x days seems to have been dropped in favour of "one Firefox installer for every language". This has its drawbacks. For instance, Firefox 1.0.3, which fixes several critical vulnerabilities, is available in English but the latest version at time of this writing in the Spanish-AR translation (South American spanish) is 1.0.2. When will 1.0.3 in Spanish be available is anybody's guess. I've just ran the firefox update checker and it tells me that my 1.0.2 is the most recent.

One then wonders, why can't one upgrade the software component side without having to care about language builds?. Why aren't "language packs" totally unaffected by minor version updates (x.y-something releases). In other words, if Mozilla.org promoted the concept of a "language pack", the version you originally installed wouldn't matter. If I'm wrong and Firefox 1.0x language packs are available, then that's not what Mozilla.org is hinting at the "other languages" download page.

In my humble opinion, it should be possible to make the installer a SINGLE EXE for all languages. It could containing the US english version binaries and all LANGUAGE PACKS together, asking the user at install time "on which language(s) do you want Firefox installed?" and then proceed to configure the language packs on top of the english binary. Heck, Sun Microsystems managed to achieve such feat with their Java2 v5.0 runtime that ships as a single, "multi-language" installer. The windows installer even detects the operating system language and launches the installation program matching the OS language. So if you're running a spanish version of windows, you get the java installation dialogues in spanish. Amazing, isn't it?.

Sun's Java2 5.0 runtime ships as a single installer for all supported languages, why can't firefox do the same by packing 'language packs'?

#2. Why deliver security fixes as "new versions", instead of patches?. Mozzarella Foundation currently forces one to download an ENTIRE 4MB+ FULL INSTALLER to get a ".01+" bug fix (the last bug fix at time of this writing, apparently affected just the javascript engine, which I'm sure comprises only a handful of files). So, why not deliver security updates as PATCHES that can be downloaded much more quickly than an entire new browser?. The technology to allow "binary difference" patches to be downloaded, applied, and the application re-started has been proved and is used daily by dozens of commercial applications.

Another related point is why increase the version number if no new functionality is added. I'm not saying that version and build numbers shouldn't be increased, but perhaps the security fixes could increase the build number rather than version number - say "Firefox 1.0 update 3" instead of "Firefox 1.03"-. It might look like a trivial issue but it's not, specially considering the the "auto downloading update patchers" issue vs. downloading and installing a new browser version on top of the old one just to get a security fix. (The download link for, say, v1.0, would always point to the latest build or update, preserving the filename, making it impossible for someone to download an out-of-date browser of that major version number, so for instance when 1.0 update 3 is released, the binaries are updated - the new users automagically get the latest code, and those running "1.0 update 2" will just fetch the updater).

Back to the issue of "update patches", and to quote a company I generally dislike but which in this instance said something clever (Symantec): "Because most software updates involve changes to only a small portion of a software application, incremental update files are generally small. The smaller data files distributed for a binary patch incremental update are often less than 1% of the size of a full update, taking advantage of the large amount of redundancy in the two versions." (*)

#3. Mozilla Foundation should offer users the option of an applications bundle, packing together the thunderbird e-mail client and the Firefox web browser, for those of us who prefer an integrated solution for Web browsing and e-mail, delivered as a single download and install. Folks at Mozzarella can call it "Firefox Suite" if they want. Then I'll be one happy man.

#4. Firefox currently is currently riding the "mindshare wave" and making headlines. So can anyone at the Foundation give me a good excuse NOT TO rename the email client I refer to as Thunderchicken to "Firefox Mailer" or such a similar name?. I don't claim to hold the truth on this one, but if Thunderchicken is the buddy program for Firefox, and it does e-mail, isn't it just logical to name it Firefox Mailer?.

Before you argue with me that "things are fine as they are right now", let me disagree: a recent article by Thomas Claburn on respected InfoWorld highlights that THUNDERBIRD is not getting nowhere as many installs as Firefox. I quote: "Firefox may be seducing users of Internet Explorer, but Thunderbird, the Mozilla Foundation's free open-source E-mail client, faces less promising prospects against Microsoft Outlook". And it mentions a report by Radicati Group Inc. saying that Thunderbird "won't pull significant market share away from Microsoft, IBM/Lotus, and Novell". In other words, even people who are not happy with Microsoft's offerings usually end up switching from IE to Firefox, but they continue using Outlook Express as their email program. This is not good, specially when the Foundation has a decent replacement in Thunderchicken, sorry, I mean, Thunderbird.

Perhaps a name change to Firefox Mailer would help?. Everyone has heard of Firefox already, a "Firefox Mail" or "Firefox Mailer" will only help, not hurt. IMHO. When I mention "Thunderbird" to the average PC user, I get blank stares, even while some of them have heard of Firefox.

#5. The Gecko Runtime Engine (GRE) was created so, in theory, all applications could share a single "runtime engine", and avoid the duplication of files. This is often not the case. If you install one version of Thunderbird, one version of Firefox, and one version of the "NVU" web authoring package -in order to obtain the same functionality that was once provided with the Mozilla Suite-, you will likely end up with several different copies of the GRE. I haven't been able to confirm what happens when you run two or three applications at once, but everything leads me to believe that there might be three copies of the GRE running simultaneously. Not very efficient, if that's the case. In fact, the old Mozilla Suite kicks the Firefox/TB/NVU combo in the teeth with regards to Gecko engine sharing. I recognize that there's some work being done here, but I'm not sure if it will fully solve the problem.. When I asked on Mozilla's IRC chat rooms late last year, they confirmed that each application installed and used its own GRE. I haven't heard about this issue since then, and INQuiring minds want to know.

#6. No splash screen. Actually this is a minor annoyance. But since I'm riding the annoyances protest wave, I decided to include it. "Splash screens suck" some powers that be seem to have decided as seen here, some time ago. And they removed one nice feature that, at least on slow systems and the first time the app is loaded, made the wait seem lower and gave a visual indication that something was going on after the application was launched. Not to mention the potential for customization and branding. See point #7 below.

#7. No "CCK". Netscape once offered their "Client Customization Kit" for their Mozilla Suite based 7.x browser. There should be a powerful, official Firefox customization tool, better if point-and-click, allowing ISPs, business and individuals that want to do so to offer customized versions, with their own custom selection of theme, bookmarks, throbber, installed extensions, even going as far as renaming the browser (as done by the Firesomething extension that afaik stopped working on the latest Firefox builds), and the like. Don't get me wrong, there's a third party tool that almost fills the gap on Windows, dubbed FFdeploy and available here, but it's not official, and the author doesn't seem to have too much time to maintain it. And it's windows-only. Not good enough.

#8. NO FTP UPLOADS. I type a ftp:// url including username and password (in the form ftp://user:pass@hostname), and then go to File->... to find there's no "File Upload" option! (I've tested this on Firefox 1.0.2 because, well, there's no Firefox 1.0.3 for "Es-AR" language, see annoyance point #1 above). Sheesh!.

Netscape Communicator 4.x introduced file upload from the ftp view in... like... 1996. That's about 9 years ago. Not adding this is unbelievable, considering that recent versions of the MOZILLA SUITE (1.7x) and Netscape 7.2 based on it, both feature ftp file upload from the File menu.

#9. BRING BACK THE TABBED SIDEBAR, at least as an option not enabled by default if you don't want to annoy current firefox advocates. This is self explanatory.

#10. Outside voices have little saying in the future direction of Firefox, or Thunderbird. Why not implement a process or programme like the Java Community Process to give the industry and user groups a voice in the future of the technology?. I'm sure this would speed up development, skyrocket the corporate embrace of the technology as an IE replacement, and remove all "fears of switching".

#11. Okay, I said ten... here's one more: An INSTANT MESSENGER (coded in xul) should be integrated. The lightweight AIM and ICQ clients on the browser's sidebar tab is one of the reasons that keep me using Netscape 7.2.

On second thought, there's an even better idea:, why not start lobbying the current instant messenger kings to do their own lightweight XUL based implementations as Firefox extensions which can be "docked" into the Firefox sidebar?.

That would certainly be the last nail on the coffin for MSIE, and a good revenue source for the Mozzarella Foundation (if there's no XUL expertise inside Yahoo, for instance, how about training them). The vision I once had was Firefox setting a level playing field for all the major official instant messenger players - imagine the user installing firefox and having a pop-up saying "you can install one or more of the following instant messenger clients on the sidebar tab, select the ones you want installed:" and below, the user seeing Yahoo Messenger, ICQ, AIM, and the like, all delivered as officially supported extensions, provided by each software vendor. (Ok, I recognize that the Vole will not be very interested in developing their own XUL based MSN Messenger for Firefox ;). But the other players have nothing to lose, as Microsoft has already played against them by making MSN Messenger ubiquitous on windows XP installations.

DISCLAIMER: this is my personal opinion, you're entitled to agree or disagree with me. If you're involved in the Firefox development and want to comment on some of my views and proposals, feel free to click on my name at the start of this article. If there are good and interesting responses we might even go as far as to print them. Since I don't want to be burned at the stake like InformationWeek's Fred Langa who dared to write his own article about the "pros and cons of Firefox" (I'm lucky he made none of my suggestions ;), be advised that flames are not going to be answered. I'll send them all to some black hole... like the MSIE team at Redmondia or the help desk for Microsoft Works.