by Michael S. Kaplan, published on 2013/11/07 06:51 -05:00, original URI: http://blogs.msdn.com/b/michkap/archive/2013/11/07/10456339.aspx
As I look across the large number of different potential future keyboards that MSKLC might eventually reach, I find myself struck by the basic convergence to two different categories that is ultimately inescapable:
The difference is striking, perhaps too striking to unfairly characterize them.
Both of them can be complicated, one by using dead keys; the other is complicated by having two-level key definitions.
So there can be a Phonetic Cherokee keyboard on the phone, for example.
It just has to be built differently on the Phone and the other 'category 2' keyboards just not with dead keys!
I shouldn't be disappointed by this, because dead keys are *hard*, and what you can do with the Phone is easier to author!
In fact, I'm tempted to say we could dump dead keys entirely, except sometimes they can produce more intuitive keyboard layouts. So they have to be an option, at least.
I just thought of something else. You know when I blogged about keyboards for other platforms in this blog? Some of the other layouts probably have a lot in common with those XML file layouts.
As I watch my own personal efforts to shape the future of MSKLC make the contrasts between these two categories become greater, I wonder what would be the best way to force MSKLC authors to deal with these two different models in a way that will be intuitive.
Perhaps two different modes, with the first category creating the same old .KLC file while the second category creates the XML files that all seem to be converging into the same format?
On the other hand, perhaps two different tools would be better, to keep layout authors from trying to use their knowledge of the first category from contaminating their authorship of the second. Would that be more intuitive?
Obviously, they should still be bundled together, shouldn't they?
Or maybe they shouldn't?
If (for example) some future version was going to run on a Windows Phone, bundling the full version with the one intended for the phone would make no sense!
But perhaps one aimed at the optimized layouts on a PC or laptop would want, perhaps even demand a bundling take place.
On the other hand, those optimized layouts would only work on the latest version, which is very un-MSKLC-like, so maybe that wouldn't be such a great use of resources, in the end.
I have time to think about it; that convergence is slow and will take a while; that chained dead keys work might be done long before that convergence finishes completely.
I should think about this more; would we ever truly need to have a version of MSKLC running on a Windows Phone, even if one was authoring keyboard layouts for Windows Phones? Just looking at the sentence on the screen makes me doubt how practical it would be, even if could ultimately prove to be useful.
Lucky for me that there is time to work through the possibilities.
What is it about the end of a product cycle that turns Program Management into Product Planning? 😏😏
They smile and say "But only one of them is actually complicated."
Doug Ewell on 7 Nov 2013 8:24 AM:
There are genuine needs for traditional physical keyboards that MSKLC can address now. We've talked about them a lot: chained dead keys, better support for characters outside the BMP. Having MSKLC read and write formats other than .klc might fit too.
Please don't let MSKLC fall victim to the growing mindset that absolutely everything has to make sense on tablets and phones and game machines, or it's not worth doing.
Michael S. Kaplan on 7 Nov 2013 9:46 AM:
@Doug, I can pretty much guarantee that it ain't never gonna happen, because the only way it could be done would be to bump up the version of the schema. This would make MSKLC incompatible with keyboards on Windows, which would truly suck!
Doug Ewell on 7 Nov 2013 12:16 PM:
OK. *whew* You should also divest yourself of the thought of dumping dead keys entirely. Users who understand dead keys (there are many, just not in the US) will never switch to long presses on physical keyboards. That's how you get repeating action.
So, do you have a rough time frame for an MSKLC revision that addresses the other items?
Michael S. Kaplan on 7 Nov 2013 12:31 PM:
I wouldn't EVER say that. I just want people to understand some keyboards do it differently and that is okay too...
Azarien on 8 Nov 2013 10:34 AM:
I just noticed that there's no longer a "genuine check" to download MSKLC.
Michael S. Kaplan on 10 Nov 2013 5:11 PM:
Do we even do that anymore?