Unicode 5.2 and OpenType?

by Michael S. Kaplan, published on 2010/10/11 07:01 -04:00, original URI: http://blogs.msdn.com/b/michkap/archive/2010/10/11/10073903.aspx

It would be interesting to learn why it is that Microsoft has got into the habit of arriving late to the party in matters Unicode.

Google search supported Unicode 5.2 rapidly, twelve months later still no sign in Bing.

Even more interesting would be any news. Is the renewed focus on standards in IE9 having a positive effect on these matters?

Ordinarily when there are many questions bundled together like this covering so many areas, I would not cover them all in one blog.

However, since pretty much all of them are outside my core areas of expertise I figure it is better not to drag it out more. I'll cover it all now. :-)

Once a long time ago I had a manager who was not entirely pleased when I suggested that Microsoft really resembled several different companies rubber-banded together more than one big company.

That image is somewhat appealing as a shortcut to describe the various business units, but probably it would be more accurate if less viscerally desirable to think of it as a bunch of metal sticks rubber-banded together, with a seemingly random set of random points where they are soldered together. Because even with entirely different business units that honestly don't care about each other some of the time, there are clearly other times that they have deliverables and dependencies that connect them.

The last question about the connection on the IE9 focus on standards could be kind of thought of in that context, I suppose. But this would not be one of those connected points. I severely doubt that one point had anything whatsoever to do with the other, with one major exception that I'll talk about in a bit. Mostly, the IE9 push for standards (e.g. HTML5) have excellent reasons that have essentially nothing whatsoever to do with a push related to OpenType (which is also a standard).

I won't comment on the IE9/Firefox issue since it's really out of my area. I do like IE9 on the machines I am running it on, though (FWIW).

In regard to Google's adoption of Unicode 5.2 versus Bing, no idea (I hardly know anyone on the Bing team and the few I do know aren't so far as I know working in this area), but if I had to guess I'd imagine that having the president of the Unicode Consortium as your "Senior International Software Architect" like Google does has to have some impact here. As far as I know, no one from the Bing team has ever even attended a Unicode UTC meeting (let alone been an officer in Unicode). The Bing team and others have to base their priorities on those old fashioned ideas (like customer scenarios and explicit customer requests) in this case, which can obviously have an impact (by contrast, the Google before Mark Davis and others moved there could not be called Unicodally Retarded, but on the spectrum from Unicode Junior Community College to Unicode Stanford/Unicode Harvard, I'd claim it was closer to the former -- probably one of the big reasons they were hired!).

And of course one cannot discount the minimal importance and meaning of support of Unicode in search compared to applications -- in that area (Google's apps) the company is not nearly as far along, even for its own applications. Mark et. al. have a big territory to cover there, it is going to take some time. I suspect they will catch up there soon enough anyway.

Anyway, none of this supposition assumes strategic thought, it is just common sense. There is a reason that Apple has supported Deseret and it's the same reaon Microsoft does so much for FarsiPersian, and it is not based merely on objective decisions of individuals with no connection to the script/language. Proximity and interest can really drive people.

"Microsoft late to the party in Unicode" is an interesting idea, but I don't see it that way; it is at best an accident of largely incompatible release schedules or at worst a lack of ability to to drive synchronization between Unicode releases and Microsoft's major product releases whwn it would make sense.

Regarding the latter idea, I'll never forget conversations that folks from Unicode had with us about how excited they were when a version of Windows was going to be shipping with Unicode 5.0 support (if only it weren't Vista, as they found out later; I imagine this revelation dampened the enthusiasm!). The fact that Windows is so common means that a new version having a new Unicode version is a tremendous opportunity to show the impact of the latest version of Unicode, and the fact that neither Microsoft norUnicode ever tries to do better at synchronization (in a legal forward-looking marketing timing kind of way, not an illegal conspiracy-in-restraint-of-trade kind of way) is actually kind of a detriment in my opinion to both Unicode and Microsoft, actually.

As for the OpenType script tags, one could actually almost think of this as a time that standards work against speedy updates, since OpenType updates mean updating an ISO standard. With the group that owns that work being a part of Windows, the simple fact that no one outside of MS was explicitly asking the standard for new script tags and the release date of Unicode 5.2 that made Windows 7 support of it impossible/impractical just makes it a lot harder to push an update to the standard when no product was going to be released in the near term.

Beyond that, the big push in the font world both inside and outside of Microsoft is not so much on adding a few script tags to OpenType; it is the ISO Open Font work in general, which touches on a lot more stuff of much wider interest. This is why I suggest how in this instance putting OpenType in a standard makes it more work to do updates, since it isn't just the time to add a section to a spec and update a web site. This is not me bad-mouthing standards, its just me pointing out that standards take time! :-)

Now these last three bullet points are not shining beacons for Microsoft either:

I guess I'm just saying that for all the complaints one could (and some do) make in general about Microsoft, in the specific case of Unicode support I think we're shooting par. :-)

Thank you for such a detailed note on the topic.

On the specific of OpenType tag names, at one level it IS as simple as placing the tags in the tag registry (www.microsoft.com/scripttags.htm) now they are formally released in IOS/Unicode standards. I've researched further and the ISO/IEC OFF document (p263) allows for this by adding to the tag registry, there is no need for a new rev. of the standard.

I am trying writing to the generic contact address for 'MS Typography' to make a formal request, although if you know of the correct email contact for tag registry I'd appreciate details (to saqqara at saqqara.org). The tag codes are already defined in ISO 15924.

Onto when the data tables hit the required spots is no more than an agility test I guess for Microsoft and others.

--------

There are some excellent points you've made which I couldn't possibly attempt to reply to. Well a couple.

I experienced first hand the way the business units work long ago. Today, 15 years on and a larger Redmond, what you describe doesn't seem all that culturally dissimilar to then. Not entirely a bad thing in my opinion although as an outsider there does appear to be a connectivity breakdown over recent years that shows in the product lineups and phenomena like the failure to fully capitalize on some great technology like .Net/CLR.

Not always late. Building NT around Unicode was a big plus for Microsoft, only slightly negated by the level of Win95 support (although nobody in 1991 expected convergence of Win 3 and NT to take 10 years so understandable).

Linking Unicode version to a new version of Windows made sense at one time, those days are long past c.f. XP upgrade pace. Likewise Office. I don't think that there can be any question among those of us who have been involved in the game for years that function/version is obsolete thinking that needs to be banished right now.

(I hardly need to comment on the merits or prognosis of a strategy of 'Bing points' paid for by reduced functionality over Google search.)