Please remove my access to the private fonts PPA

[Maia Everett]

As an Ubuntu Member, I feel uneasy having the privilege of being part of a closed beta (which I didn't even ask for) that non-members do not enjoy. I find it incompatible with my beliefs about openness and transparency - even if I'm not actually going to download from the PPA.

So, I would like my permission to download from the private PPA revoked. In case this is not possible for some bizarre reason, I would like to have my Ubuntu membership suspended until the font is publicly released (but retaining my @ubuntu.com email account for that duration, because so much of my mail is tied to it).

[Paul Sladen (sladen)]

Would it be enough to forcefully set an invalid key/token for your account's access to the PPA? This way it would be impossible for you to download from the ubuntu-font-private PPA, but wouldn't require changing your Launchpad group membership (with the risk of upsetting something in the email forwarding).

[Mark Shuttleworth (sabdfl)]

Maia

> As an Ubuntu Member, I feel uneasy having the privilege of being part of
> a closed beta (which I didn't even ask for) that non-members do not
> enjoy. I find it incompatible with my beliefs about openness and
> transparency - even if I'm not actually going to download from the PPA.

The goal in Ubuntu is to build the most effective processes for
collaboration, to get the best result. That sometimes means we don't do
the easy thing or the obvious thing. For example, we don't just hold a
referendum on every issue, we try to appoint the most talented members
of the community to make key decisions, and trust them. You could argue
that it would be more open to vote on everything, but it's certainly
more effective to run things as a firm meritocracy. Remember, openness
is a means to an end, the end being a better result. Openness is good
because it should result in a final product that meets the needs of more
people than something developed completely in seclusion. BUT, there are
useful places in between "completely closed" and "completely open", and
we try to use those places to good effect in Ubuntu.

Now, with regard to the font design process, think about it through the
lens of *effectiveness* rather than rigid openness.

Imagine if we had started with a blank sheet, and said to the whole
world (being completely open) "we want a font, what should it look
like"? The result, I hope you would agree, would be a horrible mishmash
of ideas. So obviously, starting with a completely open process from the
very beginning is actually NOT a good way to do it.

However, if we FINISH the font without ever having an open process, we
are likely to miss lots of important things that would make the font
more useful for everybody.

So the process we have settled on is one of gradually widening circles
of consultation. We start with a tight, professional design core, then
we widen the circle of consultation, as we improve the tools for
handling all the feedback we get.

The roadmap we have laid out is as follows:

 - we did the core design in a tight team lead by professionals, with me
being the voice of Ubuntu (the customer)
 - then we did an internal test within Canonical and the font design
community
 - then we made a tool for crowdsourcing font feedback (very cool, but
it had / has bugs still)
 - we then wanted to reach beyond Canonical, to a group which has global
representation and is still closely related to Ubuntu. The best group we
could think of was Ubuntu Members.
 - but we still don't want to be flooded with lots of duplicate bug
reports, and we need to start with a group that will be tolerant of
glitches in the fonttest tool, so we limit the initial feedback to that
group
 - once we are confident we have good glyph coverage and good tools, we
open the process completely

I think that's (a) a process designed to improve the font over time, (b)
a reasonable approach that achieves openness without breaking the tools
or people responsible for getting it done.

I was disappointed at your blog post, which criticised many aspects of
the process without showing any insight into that process. Did you ask
anybody close to it about it? Or did you just ...

[Tobias Wolf (towolf)]

Mark, I’m disappointed because apparently your model of widening circles didn’t include people who had enough insights in font technology used in Linux to guide the process.

What I can imagine happened in stage 1, Dalton Maag and you the customer, was that you said you wanted primarily an on-screen font with optimal legibility. So the foundry said, “yes no problem, for optimal legibility you definitely need manually crafted TrueType instructions for superhinting. That will cost you of course”. And because that sounded cool you went and commisioned TrueType instructions as well.

Unfortunately, by default Ubuntu enables the FreeType light autohinter, which has been shown to give the optimal font rendering across the board. It looks good and is highly readable with the largest portion of fonts available. It works well with the default subpixel filter that Ubuntu patches in. It works well at all point sizes and with scalable canvasses because there are no pixel snapping effects.

Ubuntu has been lauded for its superior font rendering compared to other distros. Some experts even recognize it for taking the sweet spot between Windows-style and OSX-style rendering in terms of the balancing of contrast versus shapes.

OSX has abandoned TrueType instructions long ago. People have been excited that the TT patents have expired recently, and many have said that there was a risk of litigation. There was never a risk of litigation because Apple has long abandoned the superhinting technology. Even Windows started to disregard more and more of the internal native hinting and use signal processing approaches to render fonts on pixel grids nowadays.

So why did you commission the manual hinting? It’s a waste of money. Did nobody point out that we are not dependent on this at all because we have quality font rendering in the absence of native hinting courtesy of FreeType? Switching on BCI will deteriorate the universality and robustness of our font rendering, not improve it. If it’s not switched on then commissioning it was just a waste. Look at the emerging web fonts. And the host of freely available or »free software« fonts. They all work best with our current default setup.

I’m not sure that Ubuntu Members are an inner circle of experts. Maybe they are more of a fanclub. Myself I don’t see a need to join because I can file good bug reports just fine without joining. And maybe I could have helped make the most of this project because I know a couple things about fonts and font rendering. Too late I guess.

I can add citations if anyone is interested.

[Mark Shuttleworth (sabdfl)]

Tobias, that's a valid point if your only audience is Ubuntu users who
benefit from the automatic hinting. But there are many applications
which don't, on other platforms, and the manual hinting there will help
them.

Mark

[Tobias Wolf (towolf)]

That may even make sense. I just wanted to try to point out that the »most talented members of the community to make key decisions« might exist outside of the echo chamber that is the inner circle because they are not comfortable with being part of a cabbal, even if they wholeheartedly agree with the community ideals themselves and want to see Ubuntu succeed.

But of course you’re understandably wary of letting the masses have a say in the actual font design itself. That cannot possibly work.

Only, after the design is finished and you have released the final version, I’m sure that the community at large might be able to improve it further. There are some talented people doing this already. DejaVu was enhanced a lot after the Vera family was donated by Bitstream. Mostly in terms of international script support. The same happened with Computer Modern, designed by Don Knuth, that some Polish group brought up to date and expanded greatly in the Latin Modern project. Unicode is constantly getting additions. I’m sure there will be talented people in the community that could add new glyphs in Unicode to the font, or even add entire scripts not covered yet, by taking the established design and following it. High Unicode coverage is essential in a system font. Let the community help you.

You should consider putting the font sources into Launchpad and form a project around it once Dalton Maag has released all the deliverables. Maybe after the first Ubuntu release that includes the final font.

[Maia Everett (linneris)]

Mark, I agree that the earlier blog post of mine was a little too bileful, and written in bad faith. I wrote it without knowing some facts, such as, for example, that the font was going to be released under a free license for the Maverick release. (I know also that some project leaders refuse to even discuss the question of including it into their project before that actually happens - but that's beside the point).

I'm sure Canonical, as a commercial company, has its pragmatic reasons for developing the font behind closed doors. I don't have a problem with that, per se. I'll be honest, I feel uneasy about Canonical's proprietary initiatives only tangentially related to Ubuntu the free software project (one of the first things I do after installing Ubuntu on a new machine is deleting all packages related to Ubuntu One), but: the company has done a lot in the name of free software, and not just Ubuntu, and sponsors its development at no cost for the end users; I cannot be grateful enough for that. So, if Canonical engages in now-proprietary development that will *eventually* benefit free software, that, by itself, is fine.

What I do mind is being invited into the font thing at this stage without my consent, and without even being asked. When I joined the MOTU team, and by extension Ubuntu Members, I did it with the intention of improving the free operating system that I have been using for the past five years, and sharing my contributions with the world. I knew also that non-members could submit contributions to Ubuntu like I did, they would just have to go through sponsorship.

I did not expect there to be any closed betas for components that I cannot even legally redistribute. While I recognize Canonical's right to organize development in whatever way they see fit, I find my unconditional bestowal with such unasked-for privileges incompatible with that pesky thing called "personal beliefs". A simple "Do you want to participate in the closed beta program? [Yes/No]" would suffice, and would save me (and like-minded members, if they exist) the bitterness. I believe even Blizzard, a company about as hostile to free software as one can possibly get, has the option to opt out of closed beta invitations.

[Mark Shuttleworth (sabdfl)]

 On 31/07/10 11:07, Maia Kozheva wrote:
> I did not expect there to be any closed betas for components that I
> cannot even legally redistribute. While I recognize Canonical's right to
> organize development in whatever way they see fit, I find my
> unconditional bestowal with such unasked-for privileges incompatible
> with that pesky thing called "personal beliefs". A simple "Do you want
> to participate in the closed beta program? [Yes/No]" would suffice, and
> would save me (and like-minded members, if they exist) the bitterness. I
> believe even Blizzard, a company about as hostile to free software as
> one can possibly get, has the option to opt out of closed beta
> invitations.

Maia, I'm not sure of the process, did you receive an email personally
inviting you? Or did you just become aware that the PPA was open to you
if you wanted it?

Mark

[Maia Everett (linneris)]

Below is an copy of the email I received. So yes, it personally notified me that my access to the private PPA was enabled, with no means to refuse the offer.

Subject: PPA access granted for ppa

Hello Maia Kozheva,

Launchpad: access to a private archive
--------------------------------------

Jorge O. Castro has granted you access to a private software archive
"ppa" (ppa:ubuntu-font-beta-testing/ppa), which is hosted by Launchpad
and has the following description:

None

To start downloading and using software from this archive you need to
view your access details by visiting this link:

<https://launchpad.net/~sikon/+archivesubscriptions>

You can find out more about Jorge O. Castro here:

<https://launchpad.net/~jorge>

If you'd prefer not to use software from this archive, you can safely
ignore this email. However, if you have any concerns you can contact the
Launchpad team by emailing <email address hidden>

Regards,
The Launchpad team

[Maia Everett (linneris)]

Well, admittedly, it does say "if you don't want it, ignore this". But I'm not content to just passively ignore the offer while still having the access.

[[e-mail] (mike-papersolve)]

It's 8/10 and we've heard that the beta is aimed to be open by now. Is there any new ETA on when the beta will be open? I don't believe non-members have access just yet.

[Mark Shuttleworth (sabdfl)]

We are busy getting the existing bug list under control. We are also
expressly soliciting feedback on new parts of the covered character set,
i.e. people who have a particular insight into the languages /
characters that have been added recently. There are some key matters
outstanding, like accents / umlauhts etc.

Once we have a firm handle on all of those, we will throw open the
doors. We don't want to have a flood of duplicates now, so until the
issues that have been raised so far are better understood and resolved,
we'll hold.

I believe an open team was accidentally added to the access list - that
will be dropped if it hasn't already been done.

Mark

[Mark Shuttleworth (sabdfl)]

 On 04/08/10 10:10, Maia Kozheva wrote:
> Well, admittedly, it does say "if you don't want it, ignore this". But
> I'm not content to just passively ignore the offer while still having
> the access.

Maia, this is no different than the fact that being an Ubuntu member
entitles you to an email address @ubuntu.com, and to include your blog
on planet.ubuntu.com. The facility is available if you would like to
participate. The font is supposed to represent values consistent with
those of the project, inviting the members of the project to review and
comment on it is a reasonable starting point.

 status opinion

Mark