Thunderbird 3.1
By glazou on Friday 25 June 2010, 15:29 - Mozilla - Permalink
There is something new and totally unexpected in that 3.1 I just installed on my Mac Book Pro through the update mechanism: the monospace font I'm using, I've been using for the last twenty (even 24) years reading plain/text email, Courier, is gone and Thunderbird seems to force me to switch to Menlo... I really wonder who had that excellent idea. Suddenly, my eyes are not used to the email font any more, and reading a message is less automatic.
Changing the default font is not a harmless thing. It impacts readability, accessibility, habits. It should never be done w/o asking the user. I consider this change as a strategic mistake because TB asked me nothing.
| Menlo 12 | Courier 13 |
|---|---|
 |  |
I'm immediately switching back to Courier through the Preferences panel. And I would recommend switching back TB3.1 distros to the original font.
Update: the first comment below makes me wonder if people will really understand why this is a bad change. My font choices are mine, strictly mine. The software does not have the right to force a change here w/o prompting me. Would Firefox force a change on someone's userContent.css or in the preferred background/foreground colors? Clearly no. So why is it acceptable for Thunderbird?
Comments
There were some posts about it on Planet Mozilla a few monts ago. Didn't you read these posts? I think this is right change but I'm Windows user. I cannot imagine how Thunderbird should ask user for every change it does. This would be horrible dialog/wizard.
@Pawel: the default fonts as set by a user through the preferences panel are just similar to a userContent.css. They tweak the rendering according to THEIR wishes or THEIR __needs__. Thunderbird has just no right to force a change there w/o user's agreement. Think accessibility and you'll discover this was a really bad decision.
@Pawel: "read these posts"... I probably missed it (travel, vacation, work, there are many good reasons why I could have missed them). A blog post is not a good enough reason for such a change IMO.
I agree with you. It can throw one for a loop, pretty badly.
I had to go look up the default font setup(s) on an older 3.0.5 installation to put 3.1 and put a HOWTO ticket together for the office - I suspect some of my users are going to be a little annoyed by this change too....
When I look at the time I spent to choose my font in Terminal (moving from xterm with the same font after 15 years or so), I can only sympathize with you :)
Daniel, here is the blog post talking about this change: http://monogatari.doukut.su/2010/04...
I'm not sure what the default is on OS X, did you actually modify the default setting for the font? In other words, did Thunderbird actually do anything more than to change the default pref? If it did - yes, this is evil. If not - it is really hard to distinguish the users who *chose* to use Courier from those who simply didn't care and kept the default. For the latter group it is preferable to update to the new default automatically without asking "We know that you don't care about fonts but there is this new font in town that is really much better". The collateral damage (offending the first group) cannot really be avoided then, sometimes this hard choice has to be made unfortunately...
Daniel, should I understand it that you had set specific font in TB 3 (or lower version) and version 3.1 replaced your settings (user defined setting)? Or it was replaced default settings?
I agree that blogpost isn't right channel for inform end users. Maybe release notes is right place for that but I don't think we have better place at this moment. As I said I don't think that some monster migration wizard is right for that.
I wrote the patch for this change -- no, we do not touch non-default preferences at all. As Wladimir says, there is no way to differentiate between implicit defaults and explicitly chosen defaults, and the collateral damage is unavoidable.
Also, I don't follow how the change impacts accessibility. Could you please elaborate?
@Siddarth: seriously you don't follow? Some people have issues with contrast and the way a font is designed has a strong impact on contrast, or can't read well a font that is too round (example: the menlo "l"). Their font setup is _crucial_ to them and forcing another font is just breaking their ability to read an incoming email. Most people are used to some fonts: we are all used to Courier for monospace email, Scandinavia is used to one given sans-serif font, etc. We're so used to them that cognitive researchers noticed we don't really "read" any more when the font is well known, we take a picture and that picture is immediately translasted into words w/o us having the impression of doing an effort; change the font and that's broken too.
A change that impacts an ages-old rendering style should not be allowed w/o **VERY** strong reasons to support it, like IPR on fonts. Other than that, deciding on the sole TB developers' side to impact zillions of users - including the ones like governmental orgs that often PRINT and RESCAN later - is really weird. As I said earlier, it's like changing the default font in the browser from Times to Comic Sans or tweak the default back-/foreground colors. A strict no-go.
I just don't care about the difference between predefined and user-defined prefs. I only know that making a change that a user wonders about asking "hell, what that's font, where's my usual rendering gone???" is a serious mistake.
There are tons of ways to propose such a change in an unintrusive way. For instance in the annoying "About 3.1" tab that was shown when I upgraded. "By the way, we have a new very cool monospace font we think is better for email, here's an example, click here to use it". Done.
One of the big rules of software user profiles - that I don't think Mozilla ever broke in such a visible way in the past - is "Never ever change important user settings; ever.". Font is an important user setting.
@Wladimir: I tweaked my default settings many times in the past to test some fonts and verify email rendering interoperability on both composing and recipient's side. So yes, absolutely, 100% sure, the Courier choice is mine, strictly mine, I did it personnally, that's not the default choice. And yes, it's then very very very bad.
It sounds like they're checking "Is this pref set to the default?" rather than "Has this pref ever been manually adjusted?" If the platform is capable of asking the latter question, then I think that's absolutely what they ought to have done.
But if not, I think I support their choice even though it's a bad experience for folks in your shoes. Among other things, I think it does a better job of keeping the application's look compatible with the evolving look of the OS for the typical user.
Didn't Firefox change its default fonts a release or two ago? How was that transition handled?
The word "accessibility" conjured up images of screen readers and such in my head. Even so, we felt that switching the default font to Consolas/Menlo would be a hugely better change for the vast majority of our users. These are the default fonts the OS uses, and we really should respect platform defaults. For example, Calibri/Consolas are what the newest versions of Microsoft Office and Visual Studio *automatically* upgrade you to, and I believe OS X 10.6 automatically upgrades your TextEdit font to Menlo.
And yes, you might have tweaked it, but that doesn't matter -- Courier *was* the default, and we simply do not know whether you actively chose the default or never tampered with the setting. You've worked on Mozilla for far longer than I have -- I'm sure you know that preferences are either "default" or "user", and nothing else. Courier was the "default" on Mac.
FYI, the Firefox team seems to be contemplating a similar move: see bug 468169.
"My font choices are mine, strictly mine. The software does not have the right to force a change here w/o prompting me. Would Firefox force a change on someone's userContent.css or in the preferred background/foreground colors?"
Don't be ridiculous. Your font settings were set to the defaults. You're essentially suggesting *any* UI change must first be approved by a user. Not going to happen. If you want to switch the font back, that's your choice - but your inability to recognise progress and advancements in typography shouldn't hold back these kinds of changes.
Daniel, the preferences system changed around Gecko 1.7 - the state "pref has default value but user changed it before" is no longer distinguishable from "pref was always at its default value". I think the argumentation back then was that the new approach is more consistent. But it also means that technically the two user groups can no longer be distinguished (well, they weren't fully distinguishable before either). Meaning that changing default preferences always sucks. I've done it a few times and there were always protests. But I don't think there is a better option:
1. The group of users who don't change preferences is always by far larger than those who consciously choose to use the default.
2. If somebody doesn't care about this option then showing a message is pointless, the user won't know what to do with it. If the message is obtrusive then the user will be annoyed: "What does that stupid program want from me again?" If not then he will simply ignore it.
3. The result of your suggestion would be inconsistent user experience between first-time users and those who upgraded to Thunderbird 3.1 (with sub-optimal experience for the latter group). And users have no way of knowing why. Imagine the following conversation: "Hey, your Thunderbird looks much better than mine! What did you do?" "Don't know, nothing?"
4. Finally, I think that people who do tweak preferences are in a much better position to configure Thunderbird in an optimal way after this update. If they don't like the new default they will switch back. The other group will definitely not go out and try other options to see whether there is something better than the default of Thunderbird 1.0.
I'am aware that a person who don't know what accessibility is, have possibility to change user interface of a project such Thunderbird.
Even if this a default setting, for an ugrade you can't change this, an existing default setting become an user setting. Daniel Glazman is right.
angevoyageur: Any decision to change defaults isn't black and white -- it is a judgement call (here made by the UX people). In this case, we felt that the numerous benefits to the vast majority of users (fonts that are more readable because they're designed specifically for the OS's sub-pixel rendering, keeping in step with evolving platform defaults, etc) outweigh the cost of changing the default.
A position like "this pref (which I most probably haven't touched at all) shouldn't be changed no matter how much things have improved over the last ten years" is ultimately hostile to users.
@Siddharth: "A position like "this pref (which I most probably haven't touched at all) shouldn't be changed no matter how much things have improved over the last ten years" is ultimately hostile to users."
And I dispute that. Since you can't distinguish between user prefs and predefined settings, you force changes for users who DID set fonts and THAT is hostile to them. Anyway, you seem to stand on your position but I recommend you read this tweet http://twitter.com/robinberjon/stat... ...
My sympathies are entirely with the author of this blog. LEAVE THE FONTS ALONE. There is NO legitimate reason to forcibly change them. None.
There is something here I don't understand: Siddharth (and others) said you can't distinguish between user-defined and predefined prefs. But in my TB3.1 profile, the user-defined font choices are still visible in my prefs.js. A predefined pref is not visible there... So it's still _possible_ to make a difference between predefined and user-defined: just read prefs.js