In-Cider Knowledge: FirefoxADM

The Battle of the Release Candidates

Fri, 29 Sep 2006 03:00:12 GMT

Both Firefox and Internet Explorer have hit RC1 now. Congrats to both teams - its been a long process for both of them. So, which is best?

They now look very similar (except I have absolutely no idea why Firefox and IE are displaying different content for that icon in the bottom left - I tried to refresh and even emptied cache several times).

At the moment, I am trying to use Firefox mostly at home and IE 7 at work. There is no rhyme or reason for this - just so I can get used to both of them.

Firstly, Firefox. In a previous entry I criticised several things about Firefox 2.0 Beta 2. I am still wary that this release does not represent too much of an upgrade over 1.5. However, every single criticism of the GUI is gone. Its excellent now. Small icons are in a smaller tool bar, the tabs are cleaner and match the rest of the application and the icons are well defined and excellently designed.

IE 7 RC1 is also looking very polished now. The new bookmarks location doesn't grate nearly as much as I thought it would and it is full of little surprises (I always thought that the arrow in the bookmarks panel was a visual flourish, until I realised by clicking it, it caused bookmarks to open in a new tab). The Quick tabs is hugely useful. On the downside, I have noticed that the Phishing checker is too slow and can inhibit browsing. I guarantee most will do what I did, and turn that thing off. It really needs to be more intelligent about what site it scans, not just (or seems to) scan every site.

The lack of customisability in IE 7 is really getting to me, and I'm the sort of person who never customises much. The Homepage and Refresh icons should be able to be added to the left side of the screen and taken away from the right. The security reasoning for not allowing this is a falsehood - just don't let it be scriptable from sites! As you can see, above, in Firefox, I have a design with the Google field taking up most of the menu toolbar. It works really well - allows me to see long search strings and long addresses. However, that isn't possible with IE. That's the marked difference.

Apart from that, there is very little to choose between the two. IE 7's new features are great, but matched by Firefox's robustness and extensions and themes. Whilst IE 7 will be a prominent player from early 2007 onwards, its not any win for open standards. IE 6 will still be a big player, probably bigger in market share than IE 7 until late 2007 or even 2008. And even then, IE7 does not mean sites that work in Firefox one way will work identically in IE 7 as well. To show you how bad this is, I took the University of Edinburgh's rather placid homepage and pasted Firefox 2's rendering transparently on top of IE 7's rendering:

I lined up the top left hand corners of the images. As you can see, the top left part renders OK, but as you go further down and right, IE 7 and Firefox 2's rendering gets further and further apart. With IE promising a new IE a year and Firefox busily onto version 3, web developers are going to have to support a cacophony of different versions. Reason #483 I'm glad I'm not a web developer!

More on Firefox 2.0 Beta 2

Thu, 14 Sep 2006 23:05:04 GMT

Well, first the bad news - it looks like the Disable extension button is available even to locked down users. Darn. I shall do more investigation. The upshot for the GPO Extension for Firefox is that, if this is left in, a normal (non-administrator) user could go into the Extension dialog and disable the any locked down values such as proxy, etc. This is, actually, only marginally worse than the current scenario (safe mode) but still a bind.

On the upside, FirefoxADM works on it. What?! THAT forgotten project! It hasn't been totally forgotten, trust me, but as there was a lack of requests for new functionality and I wanted to do more work on the Firefox Extension version of it, its been left to the side for, unfortunately, over a year now. I'm going to have another look at it in the run up to 2.0 to see if there is anything new to set. If you have any requests, email me or leave a comment here. Another plus is that Firefox 2.0 doesn't look hugely different to Firefox 1.5 in structure. I am definitely going to find time at work to update the WiX script to version 2.0 and hopefully add in the extra registry entries as well.

As to the browser itself, I've said it before and I'll say it again...its not got enough of anything really new that would suggest a point-zero release. For most, this might seem like a pretty small point, but trust me, it isn't! From an enterprise point of view, if you were to suggest upgrading from 1.0 to 1.5 or, say, 1.5 to 1.8, then that is just seem as a minor change. A point-zero change such as Firefox 1.5-Firefox 2.0 will be seen as a "major version change" and often will require a lot more work to justify and sign off on. Its a royal pain in the backside, especially if you understand that, actually, there was probably a whole lot more change between Firefox 1.0 and 1.5 than between 1.5 and 2.0, which both use the same branch of the underlying Gecko engine (1.8).

The theme sums up the trouble with Firefox 2. Touted as a visual refresh, it falls between 2 stools. Its not really a huge change and wont be noticable by most, but when you notice the small changes, they irritate you. For instance, you can switch between small and large icons. But all it does is make the icons smaller - the toolbar stays the same size. So, what is the point?! Is it for people who get intimidated by large icons? The other thing is the tab bar. I wish they had taken a few notes from IE 7. I really miss the little New Tab button from IE 7. Yes, you can create a new tab by double clicking in the space next to a tab, but that doesn't work too well once you've run out of space. Between the two, I'm split. Take this opinion purely as a (l)user experience, as I rarely do any web development and I KNOW that is a major flaw for IE 7 for many out there.

Firefox MSI - What a farce!

Sun, 26 Feb 2006 01:02:40 GMT

Seeing as a huge amount of people get here by typing 'Firefox 1.5 MSI' into Google, I thought I'd update the situation. Here's that Bugzilla thread again:

https://bugzilla.mozilla.org/show_bug.cgi?id=231062

Oh dear oh dear oh dear.

I actually have a lot of sympthy for that guy who just ranted madly at the Mozilla devs in that Bugzilla thread. The whole thing is a total farce. The big problem, as I see it, is that there are the Mozilla devs who have an opinion on what the MSI should be and the system administrators who know what sort of thing they need. And the two are a million miles apart. And heading in different directions.

Realistically, I think that both sides need to shut up, take a few steps back and have a long and hard think about the direction. I think that what might be a really good idea is moblising the administrator crowd and coming back with what is needed. I am considering starting a mailing list or something similar (anything but a wiki page!) dedicated to this and will mail certain people to see if I can get people onside for this.

As for me doing something, I now have a copy of Visual Studio 2005 (thanks to Microsoft who gave me a free copy of the Professional edition because I attended one of those Visual Studio/SQL Server launch conferences), have installed Votive (from the WiX project) and will try and figure out that works. At the moment, it doesn't seem to do anything. Ho hum...

Really, all I need is to be able to create a basic MSI and then, in proper hardcore MSI-er style, I'll clean it up properly just using Orca. Still, the way the MSI bug is going, I have a while to figure it all out...

Consistent Management of Firefox Across Platforms

Sun, 05 Feb 2006 01:06:37 GMT

May people who will be reasing this will already use FirefoxADM to administer and manage Firefox on their Windows platform, typically in an Active Directory. Increasingly though, an organisation will also have Linux and Mac OS X environments inside them. Typically, this involves the setting up of a seperate management setup. This unfortunately leads to inconsistent environments - and if you have a general IT policy to lock down Firefox on Windows via Active Directory, it can be difficult to copy that setup to OS X and Linux.

This post I will talk about how you can set and lock Firefox preferences on Linux and Mac OS X.

Firstly, let's refer to the key files and folders involved. On Windows, the root of Firefox is C:\Program Files\Mozilla Firefox. In order to talk about OS X and Linux, we have to also find the roots of Firefox on these platforms. On Linux, it can be many places, but is typically /usr/lib/Firefox_1.0.7 (for Firefox 1.0.7). In some distros it may be under a debug area or the like. On OS X, you can only really find this root from the terminal and will be in the apps folder. Applications tend to be "disguised folders" - however from the terminal you will find the root is /app/Firefox.app/content/MacOS/.

A key point to finding all these roots is that the folders from the root will be similar on all platforms, including an extension, default, chrome, etc folder - as you might expect from what is a cross-platform piece of software. Everything on all platforms happens from the root of Firefox and the files within.

On Windows, FirefoxADM really changes 3 files:

in the root, a file called "netscape.jsc" which does not exist by default
the file <firefoxroot>\defaults\pref\firefox.js
the file <firefoxroot>\defaults\autoconfig\prefcalls.js

In order to replicate what FirefoxADM does, you do the following:

create a blank empty file called "netscape.jsc"
append the text: pref("general.config.filename","netscape.jsc"); to firefox.js
append the values you wish to lock to prefcalls.js

So, in order to lock a homepage, you append the following text to prefcalls.js:

lockPref("browser.startup.homepage", "lockedhomepagehere");

Try running FirefoxADM on a machine and you will see this in action.

Now, using OS X and Linux, you can do the same, by using the same values, but in their corresponding location.

For instance, for Linux, in somewhat pseudocode:

touch /usr/lib/Firefox_1.0.7/netscape.jsc

echo pref("general.config.filename","netscape");>> /usr/lib/Firefox_1.0.7/defaults/pref/firefox.js

echo lockPref("browser.startup.homepage","lockedhomepagehere"); >> /usr/lib/Firefox_1.0.7/defaults/autoconfig/prefcalls.js

A similar script could be called on OS X (usually as a *nix script as opposed to AppleScript)

Obviously, if you were setting this up to run once, its OK, but would have to add your own code to make this manageable and fit into your Linux environment, and you could do more such as pipe an entire config file full of lockPref values into prefcalls.js. Some more idea would be to set this as a login hook where you throw away your prefcalls.js file every login and revert back to an original and the pipe in the lockPref values.

The scope for this goes a long way. You could make an environment which goes beyond FirefoxADM where you have a single config file that covers Windows, Linux and OS X.

Where is Mozilla's MSI for Firefox 1.5?

Mon, 12 Dec 2005 19:48:12 GMT

The answer is simple: There isn't one. I have done a bit of searching and according to the relevant Bugzilla call for the MSI, they simply dropped it from being a feature of 1.5:

https://bugzilla.mozilla.org/show_bug.cgi?id=231062

Bad, bad decision. Indeed, the more I discover about Firefox 1.5, it really makes me think that they should have developed it for another couple of months at least before releasing it.

I'm disappointed.

Talking of disappointment. Nothing angers me more than this sort of rubbish: http://www.explorerdestroyer.com/

This a low, stupid, moronic, idiotic and completely counter-productive in every single way. This isn't promoting Firefox, its making people who have it on their sites look like a$$holes - people who think they are better than others, want to restrict other, all because they use a different browser to 90% of the populace. Do me a favour.

When you see the names of the people behind explorerdestroyer, you get a picture of who they are: Holmes, Nick, Tiffiniy, and Nicholas. And no, that isn't a spelling mistake of Tiffany. That's how she has her name. They are obviously all super posh, uber rich idiots from Massachussets who have grown up being far too spoiled. Explorerdestroyer is their equivilant of them demanding toys as a kid and shouting "daddy, if you don't do what I want, I'll scweam and scweam and scweam".

Holmes, Nick, Tiffiniy, and Nicholas: You are doing the reputation of Firefox no good so, frankly, **** off.

And Here's The News In Brief...

Sat, 26 Nov 2005 02:41:27 GMT

A while since I last posted and a whole lot has been going down.

First, hello all those directed here from digg.com. Damn, there's a LOT of you. A hell of a lot more than those who were directed here from Slashdot's Rated-5-Interesting reference to this blog (which could arguable be called not very interesting!).

Secondly, thanks to Matt and his site, http://www.compteks.net. He interviewed me the other day! Very odd, but oddly, very cool. Thanks, Matt. Everyone should go and support his site. Its a great idea. It'll be the next neowin, I assure you!

Whilst on that interview, as he said "FirefoxADM has been pushed aside recently". I'm hoping to start to look at new features for it though, especially those which might fit into Firefox 1.5 (for example, SVG support). Moreover, I've been thinking in terms of what I would expect from Firefox as a corporate application. I'll start to post those ideas, and hopefully eventually get the Mozilla Corporation peoples in for their opinion. I guess some of these are things like "defining length of support for Firefox 1.0.x", "a devolvable extension management structure and the like". More in another post.

Apart from that, work continues on Firefox and Thunderbird ADMXPI. The latter is more bleeding edge than the former (due to work needs) but am currently having problems sorting out both observers for quit-application events (ie perform an action on exit of Firefox/Tbird) and profile construction issues. Tough stuff to do and there is virtually zero documentation anywhere on it.

Oh well, back to it...

The What Does What Guide To FirefoxADM!

Wed, 17 Aug 2005 21:26:18 GMT

I've uploaded to my personal web space a first-pass Features Guide to every setting inside the ADM templates. Please mail me if there are any questions about the content.

At the moment, this is a straight PDF file with no graphics. I'll bling-bling it up eventually but for now, its a dull-but-hopefully-informative 6 page document.

Can be found here: http://homepages.ed.ac.uk/mcs/FirefoxADM/FeaturesGuide.pdf

0.4 Released

Thu, 21 Jul 2005 00:07:12 GMT

Simple as that. Up on sourceforge: http://sourceforge.net/projects/firefoxadm

Remember to use it in a test structure before deploying. I think everything is looking OK with it. Certainly, it is the version which is being used live at my workplace.

Looking at 0.4.1 already. The download location is something that some people have said isn't working as it should, and I agree. I'm going to have another look at that. I may also be rejigging the ADM templates around to make them clearer. Couple of other features I'm looking at. Hopefully, the next version will be out a lot quicker than the 0.4 was after 0.3.1!

Oh, and here's a stupid note to end this post on. I can't believe it, but some wag has registered www.firefoxadm.com! Hmmm, I wonder what he's going to do with that. Certainly nothing to do with me.

Isn't the internet odd?

FirefoxADM 0.4

Sun, 17 Jul 2005 23:39:25 GMT

Yes, I know I've been quiet of late. The problem has been simple - work...and good weather. Work has been pretty hard, as it always is during the summer, but then when I leave work, I'm usually thrust into a comparitively hot summer. Woo hoo! So, I have to admit, and I know it really knocks my geek credentials down a few notches, but I've been spending time outside, cycling and relaxing when I should have been working on FirefoxADM!

But, I have been doing things here and there during work to get a couple of new things done for FirefoxADM, and a big new version is ready: FirefoxADM 0.4!

OK, the downer first. The 0.4 TEST I mentioned and linked to a couple of posts down is no longer THE feature of 0.4. This basically came down to the fact that the people at my work who were going to do some production testing of this don't seem to have had the time. As such, what I will do is include this as a script in the 0.4 package but make a point that this script is a branch of the 0.3.1 script and hopefully people can test it and it will be brought in as a feature later on.

There are 2 new features and 1 fix in 0.4 in the main scripts. The biggest new feature is about handling bookmarks. I have added in an option to redirect just the bookmark.html file to another location. For instance, at my work, we redirect them to M: drive - as such it will create a bookmark.htm file at the root of M: (and also a bookmark.bak when you subsequently launch Firefox). This is really useful for people who either (a) do not use roaming profiles or (b) do not back up their profile servers. State (b) is actually quite common and allows you to redirect the bookmarks file to the user drive which is backed up.

The second new feature is certificate handling. This allows you to copy the certificates files from a Master folder to the user's Firefox profile. The way this works is simply - set up a clean machine, install the certificates you want the users to have and then copy the relevant files (cert8.db, key3.db, secmod.db) to the Master directory on the server.

Thirdly, a fix. Basically, it seems in a mad bit of code editing, I broke the "Disable Updates for Extensions" feature. Thanks to Keith Howlett for finding this.

FirefoxADM 0.4 may well hit Sourceforge tomorrow but I want to give it a quick test first. I'd recommend befor deploying, people test these new scripts in their test setups... Anyway, early next week for the release. As soon as that is done, I want to go back to look at the default download directory setting. I've had a couple of reports it doesn't quite work as it should. It was one of the early 0.0.x features, and probably needs reworking.

One last thing. With the option in the login script to redirect the cache to Local Settings, should this be on by default?

0.4 Test Release 1

Fri, 01 Jul 2005 14:51:02 GMT

I'm just going to post up the login script for the redirected login script to my site at the moment for anyone who wants to test it. If you find any problems please email me at mark.sammons(at)ed.ac.uk.

This is just a work in progress but is for those people who are wanting to use this redirected functionality eventually.

Only thing to note is that there are two options for redirection at the moment: "Home", which is to go to a folder called Firefox in the user's AD-set home directory. The other is "Profile", which is what current functionality does. These are hard set in the top line of the login script, but will, of course, eventually be in the ADM.

Also contains the functionality for setting certificates from a master directory.

Download from: http://homepages.ed.ac.uk/mcs/FirefoxADM0.4TEST1.zip

PS: rename login script from .lgns to .vbs to use!

Of Redirection, Certificates and more

Fri, 24 Jun 2005 21:23:26 GMT

Well, Fed posted a comment on a previous post about something I was just about to post about! Cool.

I have restarted Development on ADM For Firefox.

I'm going to call this verson 0.4. That's for 2 main reasons. Firstly, with the Release Candidate nonsense going on in the 0.3 releases, I wanted to get a proper fresh start on a point release. The other reason is quite big and that is that 0.4 will have a few big changes over the previous releases, especially in the login script.

Today, at work because they are internal feature requests, I sat down and worked on 2 major areas. The first of these is certificate downloading. This works on a very similar way to the hostperm.1 and bookmarks.htm down features in the ADM, except I have built-in a choice of mandatory or only-of-user-doesn't-have-the-certificates. The second one is a major piece of surgery to FirefoxADM - profile redirection. I have had to write a chunk of code for this to make sure that any file read or write goes to and from the correct area.

By the end of the day, I have gotten both of these areas working pretty well. All I have built in at the moment is a choice between local profile or the profile being redirected to a folder named "Firefox" at the location in AD has as their Home Directory (many AD admins will know this as %homeshare%%homepath%). I will add in more but don't expect mapped drives to be able to be used - they wont resolve out for Firefox, 'fraid. Still, I would hope to be able to build in things like \\server\<username> and have ADM for Ffx work out the username (if the share has the same name as the logged in user!

Even though I have it working, I have quite a few things I want to do to it and robustly test it as well. As so much is changing, it would require a bit of a test before I'd recommend using it in a production environment.

The good thing about this functionality is that it does solve many people's issues with the profile and the bookmarks storage in Firefox. OK, so I set it up so that (a) the profile is redirecting to my test user's homedrive, and (b) enabled the functionality to "Set Cache to Local Settings". i then tested Firefox and ran it for a while. Added a ton of bookmarks and then closed Firefox. Space used on local drive. Quite a few Meg. But here's the thing - its in the Local Settings as Cache. And so wont be copied up to the network when the user logs off. The profile runs permanently from the server so nothing gets copied down at log on and log off either!

Anyway, enough of that! I probably wont be releasing this 0.4 until I have added a few more things here and there. I want 0.4 to be a good release so its gets the project back on track, and its something I will definitely be spending more time at work on as opposed to in my own time in the evening!

ADM For Thunderbird? Not yet!

Wed, 22 Jun 2005 22:51:58 GMT

It has come apparent to me that one of the major issues with Thunderbird when used in a Windows environment is that it puts all your mail files in your profile - which quickly leads to huge profiles.

I got thinking about this and started to mess with the Firefox_login.vbs login script from FirefoxADM. And came up with this:

' create shell instance
set WshShell = WScript.CreateObject("WScript.Shell")

' create file object instances
set fso = CreateObject("Scripting.FileSystemObject")

' variables I'll be needing later
Const ForReading = 1, ForWriting = 2, ForAppending = 8
Dim FirefoxProfilePath, FirefoxPrefsFile, FirefoxProfileFolder, FirefoxProfiles
Dim FirefoxFolder, PrefsFile, ThunderbirdMailFolder
Dim ParsePrefsFile, ParseOutPrefsFile
Dim FolderCreate, FirefoxProfileIniFile, FirefoxEmptyPrefsFile

on error resume next

set EnVar = Wshshell.environment("Process")
FirefoxProfilePath = EnVar("appdata") & "\Thunderbird\Profiles"
ThunderbirdMailFolder = EnVar("homeshare") & EnVar("homepath") & "Thunderbird\"

' profile creation
if fso.FolderExists(FirefoxProfilePath) = false then
if fsoFolderExists(EnVar("appdata") & "\Thunderbird") = false then
  if fsoFolderExists(EnVar("appdata")) = false then
   FolderCreate = fso.CreateFolder(EnVar("appdata"))
  end if
  FolderCreate = fso.CreateFolder(EnVar("appdata") & "\Thunderbird")
  Set FirefoxProfileIniFile = fso.CreateTextFile(Envar("appdata") & "\Thunderbird\profiles.ini")
  FirefoxProfileIniFile.WriteLine("[General]")
  FirefoxProfileIniFile.WriteLine("StartWithLastProfile=1")
  FirefoxProfileIniFile.WriteLine("")
  FirefoxProfileIniFile.WriteLine("[Profile0]")
  FirefoxProfileIniFile.WriteLine("Name=default")
  FirefoxProfileIniFile.WriteLine("IsRelative=0")
  FirefoxProfileIniFile.WriteLine("Path=" & ThunderbirdMailFolder)
  FirefoxProfileIniFile.Close
end if
end if

' check Thunderbird folder exists
If fso.FolderExists(ThunderbirdMailFolder) = false then
FolderCreate = fso.CreateFolder(ThunderbirdMailFolder)
end if

This is by no means even complete or releasible but just a proof of concept. It'll need a tidy up of the variable names for a start! What is does is, as a login script, check if a Thunderbird profile was in the roaming profile just bought down to the machine, and if there isn't one it creates it, and sets it up to go the user's Home Directory (as specified on the user account in AD) in a folder called Thunderbird. Now, the users' profile will remain small while the mail is stored on the server.

Is this the start of ADM For Thunderbird? Well, I guess it is, kinda. I'm not setting any deadlines or anything but I'm hoping that people will contact me to say what settings they would like to be able to set via ADM for Thunderbird (no locked settings though - that funfctionality is missing in Thunderbird).

Clarifications

Wed, 22 Jun 2005 22:36:51 GMT

The message last night probably needed a bit more explanation!

So, the RC pull. Basically, I set myself a level to attain with things like FirefoxADM and I was not quite sure I had it with 0.3.1. The bookmarks thing has bugged me for a while - a scratch I want to itch and recently, talking to people who find this solution not great because having the profile directory somewhere other than the default is problematic. There are other things I want to persue as well, as well as (and I know this is a bit of a WTF!! for an open source product), really decent documentation.

So, expect some stuff soon. I'm now seriously thinking of setting myself a little target. I'm always good when there are deadlines. That deadline will, hopefully, be Saturday 6th August. Why? Well, on the 8th, everything goes live where I work and the students get Firefox and our new base live. ADM For Firefox needs to be 1.0 and having no serious changes by then. That also gives me time to get all the other aspects of the project sorted out.

I genuinely feel that 0.3.1 could have been *a* 1.0 release, but not really the 1.0 release I want for this project.

And I'm going to complicate things further. Thunderbird anyone?

Next Step

Tue, 21 Jun 2005 23:23:27 GMT

Well, back from a great little jaunt on the island of Arran in Scotland last weekend - not the weekend before, as I said previously in this blog. I messed up the planning massively and arranged things for the wrong weekend. Oops! I'll upload some pics to the blog. Anyway, it was an absolutely ace time. Going to do a whole more camping from now on.

But to ADM For Firefox. 0.3.1 seems quite stable so that's cool. Some people seem to be having problems. I notice a referral URL in my blog to a forum on Coolsolutions.com where a couple of people are having problems with the ADM. If they read this, give me an email on mark.sammons(at)ed.ac.uk and I'll help you out. I think there's a bit of an order issue (ie. if you set all the GPO settings, reboot to make it apply, log in and THEN install Firefox, it wont work because the ADM is changing aspects of the Firefox that is meant to already be there!). Anyway, if anyone is having a problem with 0.3.1, feel free to mail me. I'm back in full development mode for the ADM now.

Yes, I am going to add more functionality to it before pushing for a 1.0 release, effectively it failed in my eyes the "1.0 Release Candidate" stage that 0.3.1 was put at. I presented a bit about the ADM to my work last week and some people came up with some ideas, notably one which others have asked for: a better way of syncing files such as bookmarks between network and local sites. So, what I really need to know is, what do people want from syncing. For example, for the bookmarks.htm file, where does it live, where do you copy it down to and how do you want it to handle. For example, a sync between a mapped network drive and profile, so bookmarks.htm lives on M: drive, gets copied down to profile, the original on M: is backedup to bookmarks.bak and then the bookmarks.htm file is synced back up at log off. Either mail me with your ideas or add in the comments here or on sourceforge (http://sourceforge.net/projects/firefoxadm).

I'll try and catch up with the various sourceforge comments to see if I can help there, and let me know if there are other features that are required/wanted.

Work: The Dark Foreboding Quietness

Wed, 08 Jun 2005 22:10:17 GMT

Bizarre title, but for some reason it came to mind.

See, before I went on holiday, the University where I work (University of Edinburgh) was a busy throng of students finishing up exams, saying goodbye to friends for the summer, having end of year parties. I come back to an eerie quiet. The undergraduate students are gone and as the lecturers are no marking exams, the summer schools are yet to really take off. Masters and PostGrads tend to be hard at work on end of year reports and dissertations.

Great? Not at all, as this quiet says to me: "Summer Dev". Summer is the main time when large scale development on systems takes place at Universities and other educational establishments. The goals must fit into a deadline of every big change has to be in place by the time the main body of students return (and freshers starts) in September. Well, more like August because of courses which start early.

So, I came back from holiday to find myself quickly focusing on working on a big plan, a Powerpoint presentation to the various stakeholders involved ("stakeholders": the Newest Big Buzzword!) and ultimately filling a huge whiteboard with a plan that ends up looking like the sort of thing Napoleon would have drawn up before Waterloo, with various arrows streaming in here there and everywhere. This crazy drawing, resplendant with deadlines for every Friday between the 10th June and the 19th August (apart for a couple of multi-week tasks!), was then demonstrated in full to the rest of my team, the majority of whom are senior to me. Whatever, they always seem happy to let me control the plan! For them, they probably think its the quietest option!

So, that's been taking up a whole load of my time this week.

On FirefoxADM. Ack, I mean, ADM For Firefox. No changes this week. I need to email Fed who put a comment on my last blog entry about some issues he's been having and then chase up some of the support issues there have been on the sourceforge site. I will look into that as soon as possible. The documentation is the biggest issue. I have reverted the logo on the documentation back to a modified version of the original blue logo and that will be uploaded along with a sheet describing each feature (some parts of which will be synced into the EXPLAINs in the ADM templates for 1.0). I will also try to do a FAQ answering some of the more common questions. Most of this wont be seen till early next week, as I am off on a hedonistic camping trip to a remote island off the coast of Scotland (Arran), this weekend.

Guess there's quite a few people to email tomorrow!

Back And Refreshed

Thu, 02 Jun 2005 00:53:33 GMT

Well, back from holiday and I'm feeling a whole lot better than I was before I went away. Of course, any blog readers wont know I was even away as strangely, the post I made before I went away talking about being away, the latest version of FirefoxADM and all that was mysteriously eaten by MSN Spaces. Damn and blast.

OK, well, 0.3.1 has been out there for the past couple of weeks and its looking like a surefire grounding for the 1.0. As Firefox 1.1 (aka. Deer Park) is now in Alpha, I want to get "ADM For Firefox" 1.0 done and dusted for the end of next week at the latest. Main plans are a new logo, some documentation on each individual feature, finalise the main deployment documentation PDF and anything else that comes to mind (including the project website).

I'm sorry if anyone has posted on sourceforge and hasn't had a reply from me yet. I deliberately excluded myself from computers for the past 2 weeks (apart from a couple of quick emails...) and am now starting to get myself back into the never-ending IT race. More over the next couple of days.

Phew!

Mon, 16 May 2005 21:55:02 GMT

OK, the giddyness has died down a bit. Only a few hundred referrals. Holy hell, what must a full-on slashdotting be like?!

OK, couple of things to report. No update today. There's a fix to go in but I really, really wasn't happy about it at the end of play today and got a few testing inconsistencies. The issue is simple: when upgrading between versions of Firefox, if the netscape.jsc file is left in place (because its not removed by the uninstaller), the new locked settings don't go in. I know how to fix this but just need to make sure the changes are not going to be slow or break other things. Will definitely try and get that update done tomorrow, as from Thursday (with the possible exception of Saturday), I'll be away from keyboard on a much-needed holiday!

I've posted a brief glimpse of some documentation for FirefoxADM which still uses the old, now disallowed, logo but thought I should go with it in case someone wanted to read it who came here via slashdot.

Amid a minor slashdotting...

Mon, 16 May 2005 20:41:19 GMT

Download from Sourceforge here:

http://sourceforge.net/projects/firefoxadm

Most recent documentation:

http://homepages.ed.ac.uk/mcs/FirefoxADM/ADM_Deploy.pdf

Amongst the Distractions...

Thu, 12 May 2005 01:20:18 GMT

FirefoxADM 0.3 (FirefoxADM 1.0 RC1) hit the sourceforge site on Monday. A slightly inappropriate name, given all that has happened, but don't expect ADM for Firefox to be used until 1.0.

I had a good chunk of time on Monday to make the changes and so managed to get it to release.

Very few changes in this release, but couple of big ones. Mainly, the not-to-be-supported-until-0.3 shutdown and logout scripts. The logout script had the problems I mentioned in a previous blog entry in that it didn't remove the settings applied as lockdowns, which were put into the prefs (as normal prefs, not locked prefs). I worked my way through these and that is now done. I also wanted to push on with some cookie behaviour settings as its an area which I haven't made settable. I had an idea of a pretty fine-grained approach here which allowed administrators to set lifetime policy down to the amount of days a cookie would exist for, but for some unknown reason, Firefox would only listen to the basic setting of "cookies will be: enabled, disabled, session only". So that's implemented too. The other change is the bug Jan Olbrecht brought up. I think its a pretty rare bug but needed to be catchable all the same.

This is now feature freeze for 1.0 which I expect will be in a couple of weeks (give or take a week as I'm away from a computer for at least a week and a half at the end of May). Tricky, as I have an idea for another feature, whitelisting javascript sites, but I may sit on this and other releases until after 1.0.

Any errors or feedback essential now at this stage to my account, mark(dot)sammons(at)ed.ac.uk

Thanks, as ever.

What's in a name?

Thu, 12 May 2005 01:02:52 GMT

Sorry the blog has been a bit quiet this week, but...

Terrible news.

FirefoxADM is DEAD. Finished. Gone to the great open source heaven in the sky. Over. There will never be a FirefoxADM 1.0.

Last week I was scrolling through Mozillazine.org and read a story about Firemonger, the CD-based Firefox distributable project. And then I saw something which made my heart sink: "Firemonger was previously known as FirefoxCD but the name was changed at the request of the Mozilla Foundation to meet the requirements of the Mozilla Trademark Policy". If FirefoxCD was against the requirements of the Mozilla Trademark Policy, then what hope does FirefoxADM stand?

I bit the bullet and emailed Rafael Ebron of Mozilla, who I had emailed in the past after he mentioned me in his blog. After a bit of email back and forth (and me fighting the fricking Mozilla mail server which keeps bouncing my emails!), it comes down to this:

FirefoxADM is going to have to change its name
The logo for FirefoxADM is absolutely not allowed

Damn!

The thing is, the first part, although inconvenient, I don't care too much about. You'd be surprised at the different connotations I get of the name of this project in emails! The second, I care about more so. Malte (http://malte.schwarzkopf-hh.de/blog/index.php) has done some incredible work on the logo and website for FirefoxADM and its such a shame for that to be lost, but modified Firefox logos just aren't allowed. Don't you just hate lawyers?

I'm going to review everything and start to look at the impact. The main project will stay the same name because its just a pain in the backside to suddenly rename everything on sourceforge (it'd mean starting a new project...) so that'll stay at http://sourceforge.net/projects/firefoxadm. Well, unless I suddenly get wave after wave of Mozilla lawyers on my doorstep! The worst on my side is the documentation will need to be renamed and redone (for the 58th time, it seems!). I was using a version of Malte's newest logo on the documentation and that will obviously have to be rethought.

More later on this.

But, yes, FirefoxADM is dead and will never reach 1.0.

But "ADM for Firefox", the new and approved-by-Mozilla name is very much alive! Onwards and upwards!

Horrible Decisions

Mon, 09 May 2005 00:12:49 GMT

After spending the majority of last week trying to make sure the architecture was there not just for Firefox but other applications for my big summer development push at the University of Edinburgh, I went back late last week to finish off the testing on FirefoxADM 0.3 and to release.

And Woah! I found a problem. The logout script, only to be supported from 0.3 onwards only removes those settings set by login script but not startup script. A real oversight on my behalf. I had a very difficult decision but decided to not go with the release. Expect 0.3 early this following week though. Its my main priority.

If you have been following the headlines, there is a critical remote execution bug in Firefox 1.0.3 (and, I presume, all older versions). I suggest that if you are using FirefoxADM, you enable the "Disable XPI installs" function until 1.0.4 is released (and let's hope it is...soon!).

Creeping features

Tue, 03 May 2005 20:51:14 GMT

Not quite to 0.3 (1.0 RC1). I have actually been working on other parts of the project the last few days including starting to get some proper documentation done as well as, somewhat most importantly, getting it ready for a 6-7000 computer install. Tis done anyway.

The fix is in place now, so that should be everything for 0.3. Except, I couldn't let it lie, and I added another feature in - proper cookie handling. FirefoxADM will now allow administrators to decide whether to throw away cookies at end of session, keep for a certain time period or whatever. More on that when I test it properly.

I'm still looking at a bit more accessibility functionality but am finding very little evidence of people having used things like Text-To-Speech within Firefox. Disappointing in I can't get it quite to work, but there is evidence that working with the "caret" you could use screen readers. Will look more into that.

And one other thing. It actually looks as if there is no locking functionality available for Thunderbird. Make of that sentence what you will...

0.3 (1.0 RC1) Delayed....Slightly

Sat, 30 Apr 2005 01:40:16 GMT

A bit of a problem rose late this week and therefore, I'll need to add the fix in before I release 0.3 (AKA 1.0 RC1).

Thanks to Jan Olbrecht for finding the problem (inconveniently for him!). Basically the problem is that the "Use IE's settings" for proxy in FirefoxADM Login assumes that there is a file to read, but if you are using the redirection for things like Application Data, IE's settings are magically moved to Local Settings\Application Data\... As ever, random stuff from Microsoft's Group Policy redirection settings. Its pretty easy to fix but I'll need to test on Monday. Expect 0.3 then. However, it is very likely that there will be no changes between 0.2.3 and 0.3 apart from supporting the 0.3 scripts.

On the 1.0 issue, there are already things popping up for which I am not happy to go with pre-1.0 like handling bookmarks and a different method of locking. These, if they happen, will likely occur post-1.0. I should reiterate that 1.0 doesn't signal the end of development for FirefoxADM, just it will be in a different state of development and will be supported by myself. Talking of which, I'm getting emails at least daily from people about FirefoxADM. That's great. I'm glad its helping people. It was brought to my attention that maybe I should look into quotable Success Stories as a method of marketing. I like the notion of FirefoxADM spreading by word of mouth but maybe its something. If you have a "Success Story" I could possibly quote on the FirefoxADM site, that would be great. Email me at mark.sammons(at symbol)ed.ac.uk

One thing that kinda made me interested was discovering the Statistics part of Spaces a couple of weeks ago. Wow! There's a whole lot of people getting here via Google and MSN Search! And some of the referring search queries are really interesting. I probably should add the place to download FirefoxADM to this place more, if not with each post!

So here goes:

To download FirefoxADM, go to http://sourceforge.net/projects/firefoxadm

Apparently...

Fri, 29 Apr 2005 00:03:40 GMT

At a session at WinHEC, Microsoft said corporations wouldn't use Firefox because they can't use Group Policy to deploy it and lock it down.

Sheesh, I really need for FirefoxADM to become more well-known amongst the masses out there!

That's why I did the project. Not Microsoft bashing at all, but to allow Firefox to be controlled by group policy and locked down.

All Aboard The Last Train To Releaseville

Wed, 27 Apr 2005 22:47:13 GMT

OK, A quick freshen up on FirefoxADM.

This is really the last call for features before 1.0. If you have any requests, email them to mark.sammons at ed.ac.uk or put them in the requests on http://sourceforge.net/projects/firefoxadm.

The next stage is a little bit boring - I'm going to be doing some documentation, some write ups, work with Malte (well, approving what he does!) on the http://firefoxadm.sourceforge.net site and generally spruce up the project in time for a 1.0 launch. I expect the 1.0 to hit around mid-to-late-May, given everyone's workloads. I really then want to push it because that is the time when Universities and Schools start to consider what software they will use in the following 05-06 sessions. I want to be able to aim the project at them (and if you work for a school or Uni and have good reviews of FirefoxADM, please pass on the information to colleagues in other institutions! :-) ).

I do have several ideas brewing for post-1.0, but more on that another time!

Its full steam ahead!

The Numbers Game

Tue, 19 Apr 2005 21:52:26 GMT

In the past week, I have done very little on the code side of FirefoxADM, or even on the documentation. The past week has really been about adding that final toughness by hardcore testing to FirefoxADM, and I'm happy to say that everything seems fine and of the sort of reliability I had hoped for.

Futher to that, there have been a couple of external people I have been helping in my own time to just try and make sure they understand everything. The reasoning for this is that we are talking big number "clients".

Ahhh, yes, numbers. I really am not into the whole Numbers Games, and by that I mean, FirefoxADM is on x workstations. I ask of no registration or feedback on numbers or the people using FirefoxADM because, frankly, its none of my business how many machines people are using on it. Also, doesn't really matter to the project because, by design, "if it works for 1 computer, it should work for 10,000".

The downside of this is that I never expected FirefoxADM to be in the position it is now. The numbers don't matter in terms of the coding and privacy sides, but at the same time, I want to get the word about the project out, to give it some real credence. I would hate for people to come to here or sourceforge and think "yeah, but I bet no one is using it in real life", when in fact there are a LOT of people using it.

So, a quandry. I decided I don't want to be as silly or go as far as request people email me to say how many computers they are using it on, if I haven't already talked to them. And I don't want to just leave it hanging here.

So...

I'll say this, and yes, its deliberately obscure. What I know is that FirefoxADM is being used on a number of machines that is high into 5-figures.

Is that number the majority? I can't tell for sure, but at a guess, hearing where people found out about FirefoxADM from, those who haven't told me how many machines they are using it on, narcissistic searches on the internet and the amount of people who suddenly came out of the woodwork when that obscure Windows 2003 SP1 bug appeared, I very much doubt it. I wouldn't be surprised if the real number was much larger than the machines I KNOW about. Quite probably a lot more.

I know this might sound stupid, but I hated writing that paragraph above! I'm British and anything that seems like boasting makes my skin crawl. That's just the way of us Britishers.

The thing is, though, as summer comes tumbling ever faster, the relevance of FirefoxADM will pick up. This will especially be the case for Academic Institutions. Remember, for the majority, this is the first window of opportunity to install Firefox and many will go for it - schools and universities tend to only make big application changes during the summer period. The locking and setting of preferences is a MASSIVE requirement for most academic IT departments (as well as not storing the cache on the network, to be fixed in Firefox 1.1, but already fixed in FirefoxADM 0.2.3!) and without them, many will not be able to go with Firefox in their deployments. Combined with a decent MSI such as the one found on www.frontmotion.com, FirefoxADM allows this, and this has to be one of the major areas to "market" to post FirefoxADM 1.0 and into the summer.

Anyway, I hope the above doesn't sound mental or big-headed. It isn't meant to be! Over-enthusiastic, quite probably - though!

0.2.3 and Beyond...

Wed, 13 Apr 2005 18:56:47 GMT

A couple of days ago, I uploaded FirefoxADM 0.2.3 to sourceforge. There was one little difference to the one I attached to the last entry. Unfortunately, all it came down to was forgetting to add quotation marks to the Local Settings cache functionality, but it was a crucial bug which caused the entire of 0.2.3 to fail. Anyway, that's fixed so don't download the one below, get the proper one from http://sourceforge.net/projects/firefoxadm.

Change log of 0.2.3 was:

- Fixed: Anomaly on Windows 2003 SP1 that fails to load FirefoxADM Login ADM template ("Found: CATEGOR, Expected: CATEGORY" error)
- Added: Ability to enforce SSL disabled in Cache, in FirefoxADM Startup
- Added: Use Local Settings for Cache, in FirefoxADM Login
- Added: Ability to disable Prefetching, in FirefoxADM Startup
- Added: Ability to enable or disable local URL from network (file://), in FirefoxADM Startup
- Added: Ability to set whitelist for Integrated Authentication sites, in FirefoxADM Startup
- Added: Ability to set browse with keyboard "caret", in FirefoxADM Startup & FirefoxADM Login

A quick rundown of these. The Fix was the END CATEGOR bug that appeared somehow when using FirefoxADM Login's ADM template on Windows 2003 SP1. I still have no idea why it does that, except possibly the return acts as a confirmation of the line. As such, any version which displays this can be fixed with a carriage return at the end of that final line. Maybe it was built in by Microsoft to deliberately destroy the template! In which case, I'll threaten to sue, and get paid off with a load of money! Hey, if it works for those morons at Real Networks...

New features? Well, the most significant is that FirefoxADM Login has functionality to set the local cache to be in C:\Documents and Settings\<username>\Local Settings, and thus not be copied up with the profile. I, unfortunately, could not implement this as a Locked setting because it deals with a user's profile, but this is the next best thing (until it is hopefully fixed in Firefox 1.1). I will probably add more to this functionality to offer the ability to remove any cache that is in the roaming profile, to help those administrators when moving across to the new functionality. The other big addition is the ability to set a whitelist for Integrated Authentication sites. Again, I may improve this so that it can also be a per-user setting, but a locked setting (per computer) for now. Several of the others are just tightening things down. In response to Google saying they are going to user Prefetching in Firefox, administrators who don't want them to can disable this, I've extended the cache-ing functionality to stop people stupidly cache-ing SSL documents, I've added in functionality (a contribution by Michael Morandi) to allow or disallow usage of the file:// protocol, and started to add in more accessibility functionality by allowing the usage of Browse with caret...basically adding more functionality to browse with keyboard (as a per machine or per user setting).

From here, the next version, with all feature requests I know of (if you have more, let me know!) handled, will be 0.3. This will unofficially also be 1.0 RC1. The reasoning for this is that:

(a) the project is getting mature enough to the point where there aren't that many major to-dos on the coding side,

(b) with the amount of usage of this getting into the "good grief!" numbers (more on that later!) and that usage being very much in production environments, it seems daft to say "well, don't rely on it - its only a beta product"

(d) It would be great to be able to hit 1.0 in a timely manner so that I can start to sync FirefoxADM version numbers with Firefox's, and thus work on having FirefoxADM 1.1 done so that it takes advantage of new functionality in Firefox 1.1 (whatever that may be).

Well, that's all for now. I'm off out to enjoy a rare warm Edinburgh evening...

Turning FirefoxADM Back On

Mon, 11 Apr 2005 21:48:58 GMT

As I said previously, FirefoxADM has been going through a quiet period whilst I worked on other areas. Fortunately, today marked the end of a major phase of us rolling out Service Pack 2 (a major part of my job over past months) and can now look to other areas, even as far forward as September and Academic Year 2005-2006.

In that fits Firefox, and thus FirefoxADM. What's really good about this is that I am now able to enter a relatively quieter period where I can actually settle down and do some hardcore development towards future versions of FirefoxADM, with the goal of the mythical version 1.0. Over the next week, Firefox will be a major area I am going to concentrate on, so if there are feature requests, get them in now! The next version of FirefoxADM, 0.2.3 will almost certainly be released tomorrow. There are some big added features here. Firstly, it now supports the cache being redirected to the Local Settings in the profile, quite a big bugbear, I know. Added to the cache, I now allow administrators to prohibit that SSL will be cached. As previously mentioned, I also wanted to disallow the Prefetching functionality that Google have tapped into.

Also, there is now special functionality to allow file:// to be handled (disabled and enabled). I also have added in more Accessibility functionality to allow the "Browse by caret" to be set as a per machine or per user setting (locked, or default). Lastly, and not least, I have enabled a white list of Integrated Authentication (NTLM/Kerberos) sites.

As ever, all this could change if they don't test out properly. However, if you wish to test (do not deploy!), it can be found at http://homepages.ed.ac.uk/mcs/FirefoxADM_0.2.3.zip

In other news, the web site is starting to get a nicely professional look thanks to the super talented Malte (http://firefoxadm.sourceforge.net). Hopefully, more on that including a change of that temporary logo to something a little bit more unique. I am hoping that once all this is in place, I can also add to future versions of FirefoxADM with usage/deployment white papers and the like.

Lastly, a bit of an odd bug has sprung up in the past week or so. People deploying to Windows 2003 SP1 are finding that the firefoxdefaults.adm file refuses to load with the error:

Found: "CATEGOR", Expected: "CATEGORY"

- even when copying the file straight from a domain where it was working. The file looks fine but I have been able to replicate this on a 2003 SP1 RC2 box. The resolution? Add a carriage return to the end of the file. Very, very odd. I'm going to keep an eye on this one...

Quiet Before The Storm

Sun, 03 Apr 2005 02:00:23 GMT

Its another patent-pending In Cider Knowledge Super Late In From The Pub And Not Tired Blog Postings!

The reason that there hasn't been a post on the blog this past week is that I have put FirefoxADM slightly on the back burner over the past week while the paid work takes precedence. See, its Easter holidays and the students have been away over the past week and will be the next 2 weeks. As such, this is the perfect time to do those Big Changes without its affecting too many students. Unfortunately, due to leading the development on our Supported XP Desktop's SP2 upgrade, and it not going as smoothly as hoped, its been a very busy one. The problem is that I set a couple of deadlines a couple of months back for 4th April (Monday) and 11th April (a week Monday) to do a couple of major changes. I'm totally unwilling to carry on having this SP2 monkey on my back dominating everything I do and so am going to make damned sure I hit those deadlines.

However, there is good news. My work, The University of Edinburgh is looking to move from a combination of Mozilla 1.7 and IE 6 as its supported browsers to Firefox 1.0.2 and IE 6 in the Supported XP Desktop, and FirefoxADM is a big part of that decision. We are talking, easily, 6000+ and its ever growing. The big point of this though is that is this will mean that I will get real time to develop FirefoxADM at work. I plan as soon as I can after April 11th to sit down for probably a couple of weeks, prep a new MSI for Firefox and then work hard on getting as much as I can into FirefoxADM as possible. I hope I can dedicate at least 2 weeks to the project (76 hours of work time or so!) and really polish it off. This also gives me something to aim for not just for the 0.3, but the 1.0. More on this another time.

Well, enough disclosing of policies at work, there's more interesting stuff to come. Malte Schwarzkopf, who has that all too rare thing: an interesting blog (http://malte.schwarzkopf-hh.de/blog/), and myself have been in email contact over the past week regards doing a unified web site for FirefoxADM and the other possible offshoots. Offshoots? Well, as well as MozillaADM (should I do this? Let me know if there is a desire for it), there could be a Thunderbird equivilant as well. I'm not developing that myself at the moment so I wont go into that. I'd really love to take the whole *ADM idea generic at some point as well, writing something that allows you to generate your own applicationADM. That's a bit of a pipe dream at the moment though.

The goal for 0.2.3 is getting concrete at the moment. I've already enhanced the Cache abilities to allow the administrator to deny that SSL is cached. Also, currently, the cache is saved with the profile. Bad, terrible thing. Its the subject of a 4 year old (its birthday was last week!) bug on Bugzilla. This is now fixed in FirefoxADM Login (subject to testing, but running for a week fine on my local machine) so the cache is now saved to the non-roaming Local Settings are. The rest are todo's. A "Might Do" is to enhance the individual settings for Type Ahead Find. The rest are definites. As an accessibility issue, I am going to either lock or set as a user setting the "Browse With Caret" option, a feature request on SourceForge to pass-through NTLM page security, a feature already written by a Swiss administrator to allow or disallow file:// protocol and, in light of stories this week, allow the administrator to block Firefox's Prefetch functionality which Google have decided they are going to use - y'know you happen to want to decide what to do with your bandwidth and not have Google decide (I'm really growing to dislike that company). Much to do, and I hope to do things here and there this week to get to a state where I can clean up at the start of next week, test and release. Its possible some things might be added or dropped off this list.

Anyway, to bed...

FirefoxADM 0.2.2 And Other Stories

Sun, 27 Mar 2005 08:54:05 GMT

Its British Summer Time! Despite that, a very early morning. Up after an evening of watching England fans celebrate their football (that's proper footie, ie. soccer, as known to you on other side of the pond) hammering of Northern Ireland, and Scotland falling to yet another loss.

I finally got FirefoxADM 0.2.2 released to sourceforge (http://sourceforge.net/projects/firefoxadm). Changelog was as per last post. Testing was fun. Its amazing just how locked down you can get Firefox when you want to. Everything seems to be fine on that side of things, and the login script issue I mentioned before seems to be connected to Netware. Stupid, horrible, please-die-soon Netware.

The list of things-to-do for FirefoxADM 0.2.3 is growing though. The one real biggie that is getting to me now is this: https://bugzilla.mozilla.org/show_bug.cgi?id=74085. I really, really, really, want to fix this. Its a stupid, seemingly easy to fix, bug. I'm not only certain that I have the architecture built into FirefoxADM Login for this, but I'm thinking additionally making it a seperate script which any administrator can use. I wont say any more because, knowing my luck, this wont be as easy to fix as I think, but I agree with the guy who brought this to my attention, Christopher, this is a total showstopper for Firefox in the enterprise. Anyway, more on 0.2.3 after the weekend.

In my opinion though, this has been a pretty terrible week for Firefox. The biggest personal blow was the, frankly, ill-thought decision for Firefox to stop supplying ZIPs of official builds. I find this a totally incredulous decision, given that (a) the Zip builds are often used in environments where the installers are not applicable (as in portable Firefox, admins who want to copy the zip extracts to the network and run them from there, etc etc etc) and (b) the official installers are such rubbish (and yes, I am including the current state of the MSI build for Firefox). Personally, this effectively kills an idea I had which I had toyed with out of sight and mind of this blog. The idea was a way of using the Zip builds to have an enterprise-size autoupdate solution. It seemed to be a real solution. Now, it is dead and I wont continue with this idea. I'm really pissed off with the situation - its a real 2 steps back decision.

The other was an interview in that IT tabloid, slashdot, by Mitchell of Mozilla. She really had a bad one, in my opinion. The worst part came from a potentially difficult question from someone questioning the fact that a bug which wiped people's hard drives had been unfixed for over a year. Her response? Brushed it off as if it was meaningless. I couldn't believe it. Further investigation into the bug saw that it was caused if a user installed Firefox to somewhere like C: or C:\Program Files\. Uninstall, and it would deltree everything underneath the install directory. So, install to the root of C:...uninstall and lose everything on the C: drive. A nonsensical piece of logic from the installer/uninstaller, as if you should only ever uninstall what you install, but.... Here's the thing. Read the bugzilla thread on this and what does it come down to? Mike Conner saying its no problem on the installer but a problem of the users who are stupid enough to install it to the root of C: or whereever. ARRRRRRGGGGGHHHHH! No no no no no. A daft, arrogant, ridiculous response. It'd be the sort of reply to suggest Mike steps away from Firefox's development for a bit, if it wasn't for the fact that in Conner's own blog, he said he was the noly one of the 6 main developers of Firefox doing anything. Christ.

Well, that's that off my chest. I totally believe Firefox needs these kickings to make it better. Its often said that Microsoft is where it is today because its a big gun who fights like the underdog. I totally believe that. If Firefox wants to succeed, it needs to fight like that, stop thinking its better than IE in every way, stop selling Firefox in that way (and that means a radically shake-up of the cripple-pages-viewed-in-IE pish "marketing ideas" put up on Spreadfirefox.com) and fight harder and more intelligentally.

More later...

Coming In For Another Attempt

Wed, 23 Mar 2005 22:56:31 GMT

0.2.2 will release tomorrow, hopefully.

The delay since the last post has been coding in extra stuff, a weekend in which I was stupidly busy helping friends and then ill Monday (stupid curry!). Last couple of days have been catching up at work, as well as trying to overcome a bit of a temperature that came with the Monday illness.

Current release change log is:

- Fixed: FirefoxADM Wipeout renamed FirefoxADM Shutdown (0.3 Script)
- Added: Ability to set whitelist for allowed popup sites and install sites by list in group policy, in FirefoxADM Login
- Added: FirefoxADM Logout script for removing settings that were added at login (full Group Policy emulation) (0.3 Script)
- Added: Ability to disable javascript, in FirefoxADM Startup
- Added: Ability to disable SSL2, SSL3 & TLS, in FirefoxADM Startup
- Added: Ability to disable IDNs, in FirefoxADM Startup
- Added: Ability to disable password remembering, in FirefoxADM Startup
- Added: Ability to enable Type Ahead Find, in FirefoxADM Startup & in FirefoxADM Login

There are a couple of other things which I have decided that I wont put in until 0.2.3 (not 0.3, yet!), including extending the functionality of the cache and type ahead find, looking into a feature request that has been posted on sourceforge, looking at ways of clearing the cache on logout or, maybe better, redirecting it to a local settings temp directory within the profile.

However, the plan with 0.2.2, combined with some functionality in 0.2.1, was, as I posted before, to be able to allow a really, really locked down Firefox. I really want to do a "Scenario" document on rolling out Firefox and be able to say to administrators to lock down Firefox and release early, and then be able to ease off the restrictions if they want. However, more likely is the ability to have as much control and ability to restrict as possible.

See, my belief is this. In an enterprise, if you deploy Firefox, IE will still exist on the machines. As such, an enterprise needs to have a strategy for both - and the best strategy may well be a case where IE and Firefox co-exist but have different rolls or they may co-exist and via Group Policy, administrators can change the set ups of the browsers to offer the best resilience. For example, if IE had a really bad virus going round the internet, the administrator has the option to close off IE's internet access via security zones or proxy, yet can continue to allow the internet "service" to continue via Firefox. Alternatively, maybe a company has an image of Firefox 1.0, cannot put in 1.0.1 and want to stop the IDN issue. This ADM allows that.

This all leads to 0.3. What will that be? That will be the first Release edition. When I say that, what I mean is, a lot of documentation, scenario documents, etc, will be based on 0.3. As such, by 0.3 I want to get as much functionality in and be as fully-featured as is required by administrators.

So, that's all upbeat. I'll end on a lower note, therefore! I will be interested to hear from anyone who isn't quite getting the login script side of this working. I am having mixed results with it sometimes not seeming to pick up any environment variables, which I think is possibly connected to my work AD's strange script settings, so might not be affecting others.

Anyway, thanks for everyone who has sent me comments and massive apologies to those I haven't emailed yet (especially Malte!). I'll get to you soon enough.

Landing Aborted

Fri, 18 Mar 2005 02:16:36 GMT

0.2.2 will not land tomorrow. Too much work to do at work, I'm afraid.

I've decided to roll up 0.2.2 and some of the stuff I was thinking for 0.2.3 into a bigger 0.2.2 release next week. A lot of the updates in 0.2.2 will be about securing and disabling functionality and just stopping Firefox doing stuff. This is kind of what I was thinking about as a way of smoothly implementing Firefox into your environment on a least-priviledges basis (ie. allowing Firefox but with only functionality you trust it with and then gradually easing off those restrictions, possibly finally with Firefox getting into a mainstream position and so you can start to close off IE!).

More news over the weekend...

Permission to Land

Wed, 16 Mar 2005 23:24:00 GMT

I really should stop reading Mozilla pages. I have now got into my head to use "land" as opposed to fix a bug/add a feature. Stupid Mozilla meeting notes.

Still no 0.2.2 released. The reason why is that over the last few days I have had my head buried in XP Service Pack 2. Here's the thing, machines in a lab environment, locked down, with proxy server settings (even if not needed). Launch Internet Explorer in SP2, and it starts off, goes to our internal home page ok. Type in external URL and IE hangs completely and has to be killed after "Not Responding" (although occasionally, it will get as far as a few pages in before hanging). Get rid of the proxy and its fine. Proxy fault? Can't see anything wrong happening on the network. I have literally been looking at this one issue for 2 days solid. I find it today and the problem appears to be an obscure group policy setting called "Make proxy settings per machine". Well, seems like it. Finally landed that one with help of colleague and need to make sure and test tomorrow. After that, and assuming this is the problem, I'll need to hack my way through the behemoth of Microsoft's site to find how I can submit a bug in SP2...

Talking of bugs, I was contacted and asked to add my comments to this bugzilla bug #267888, which is an attempt to get group policy handled within Firefox. An admirable attempt, to be sure. I am so sceptical about this being possible even in the medium term, that whilst I will monitor and add to the bug if possible, I don't hold out too much hope. I personally would love to go and hack at the source code, but I believe that until this project is in a really, really stable release state and I feel I can't add too much more to it (ie. the mythical 1.0?!), I'll concentrate on FirefoxADM. Also, plans are not only afoot for FirefoxADM, but I'm starting to look at how I can branch FirefoxADM to make MozillaADM (SeamonkeyADM?). And, well, as Yoda said to Ben in Empire Strikes Back, "there is another". I'll leave it at that!

Due to workload, I am now doubting I can get 0.2.2 up to sourceforge before Friday. I'll try. Plans for 0.3 are starting to get more concrete though. Testing of the logout and shutdown scripts is the big one, as is incorporating a slight change to the way the code handles settings. Lastly, and least, I need to start to look at how to build firefox.sourceforge.net as this project's homepage. That one might not happen due to my unbelievably terrible art and design skills!

Creating an image containing Firefox with locked down settings

Sun, 13 Mar 2005 00:13:57 GMT

This one came to me earlier on this evening. Some people, nay, a lot of people, will be deploying Firefox as part of an image but don't know how to do a locked down Firefox in that.

Whether you use Active Directory or not, you can use FirefoxADM with the local group policy to lock down Firefox before creating an image of Firefox.

I'll do a tutorial on this, hopefully next week.

Saturday Night's All Right (For Coding?!)

Sat, 12 Mar 2005 23:47:50 GMT

Its an unfortunate Saturday night in this week because money's too tight to mention (alright, enough crap song references!). Working for a University has a few perks, but a number of downsides - money being a big one. Unfortunately, its not just money because Universities can't pay, but because they act like an anti-Meritocracy: You don't get paid on how good you are, but how long you've been there. Anyway, enough of that...

I'm hoping to release 0.2.2 early next week. In the expected change log so far:

Back to the 0.3 scripts in a minute. First the other additions.

The main was being able to define a white list of allowed popup and install sites. This actually will replace the "Set hostperm.1 master file" from 0.2.1. I'll leave the functionality and code in 0.2.2 for that, but it will be marked "do not use", deprecated and removed in 0.3. The new functionality allows you to set a list of pop up and allowed install sites in a ";" delimited list, and each setting can be marked as "NONE" to remove all.

The others are more locking down settings. Seemed a bit silly to offer the ability to disable java if javascript wasn't also able to be locked down. The other is SSL, SSL2 and TLS locking. That's for the ultra-cautious admins. At the moment, that's an individual on/off switch which disables all 3 options. If the ability to turn them off individually is required, request it...

The 0.3 scripts are the logout and shutdown scripts for FirefoxADM. These are for emulating fully managed Group Policies so any settings aren't tattooed if a group policy is removed or the user goes to another machine where no settings are autoset. They wont be required at 0.3 but recommended if you are using a fully managed environment. They are also usable for odd situations. Like, for instance, you have used locked settings on all your machines, but your Lab Admins who have rights over the machines don't want locking. You can run create a group policy with security filtering of those Lab Admins and, and this is where it gets odd, set the shutdown script to run as a login script and startup script to run as a logout script (ie. Startup locks settings, Shutdown removes locks - so at logon, these guys would get a logon script of unlock settings and a logout script which locks the settings up in time for the next user).

For this reason, I'm seriously considering how I can rename the startup, shutdown, logon and logout scripts to lock, unlock, setpref and unsetpref respectively. The only problem will be how to teach users that in normal practise, a lock script goes in at startup, etc etc etc. Any ideas, mail me...

Anyway, these are the first parts of what will become 0.3. That's not to say 0.3 is the next version. It might not be. Indeed, I've got an idea of what I'm aiming at for 0.2.3 already. More on that after 0.2.2 is up on sourceforge.