2011/12/21 Carsten Haitzler <[email protected]>:
> On Wed, 21 Dec 2011 11:44:29 +0900 Bluezery <[email protected]> said:
>
>> 2011/12/20 Tom Hacohen <[email protected]>:
>> > On 20/12/11 12:27, Carsten Haitzler (The Rasterman) wrote:
>> >> agreed. programmer should simply try and use pinch gestures and if he
>> >> doesnt get any either its not possible (no multi-touch device) or, no
>> >> emulated gesture/input is being used (he really is looking at zoom not
>> >> pinch so you can emulate with mousehweels etc. or other actions like
>> >> "double tap" and so on). the programmer shouldn't decide by system
>> >> configuration decides - if someone wishes to violate the patent by turning
>> >> it on, they can, but otherwise it's off by default. in theory we maybe
>> >> should also shoudl be able to compile out the code to handle the bevhavior
>> >> which at the app level will result in the same thing as it being off in
>> >> config - it just can never be turned on without a recompile (freetype did
>> >> this with bytecode hinting for many years).
>> >
>> > Yep, freetype is the same example I had in mind.
>> >
>>
>> Thank you for your comments. :)
>> Besides, I have a question.
>> I know that bytecode interpreter was patented previously and It was
>> disabled at compile time in Freetype.
>> If so, bytecode interpreter was not included in released Freetype binary???
>> But elm_gesutre is built in when elementary binary is released and it
>> can be disabled at runtime time.
>
> well freetype offered just the compile-time switch. runtime the app requested
> which kind of hinting they wanted via an api. as such efl always requested
> bytecode, BUT if freetype didn't have bytecode hinting it would fall back i
> think to autohinting. that patent has now expired so it's not relevant 
> anymore,
> but taking it as an example many linux distributions disabled bytecode hinting
> entirely. others kept it enabled BUt configured everything by default to not
> use it (use auto hinting). this seems to have worked quite well so far.
>
> another thing that came to mind is... make the behavior pluggable. yes - it's 
> a
> "workaround" but this is the same workaround linux distros use for mp3. don't
> ship with any mp3 codecs, but users can add them later themselves from "3rd
> party packages". as such the vendor doesnt violate any patent. someone making 
> a
> module to recognize pinch "events" and pass out some zoom events is probably
> not violating a patent as the patent probably covers a much wider pipeline and
> such a module is simply translating/interpreting events. even if it violates,
> it is just this 3rd party module that violates, not the system that it plugs
> into.
>
> so my advice to do this best is:
>
> 1. allow it to be compiled out OR even better - make it a module.
> 2. if you make a module, make sure that it can be built totally separately as 
> a
> stand-alone piece of software.
> 3. allow the module configuration to be able to bring it in and then enable 
> the
> feature *IF* a user installs the module and changes that config
> 4. leave it to "3rd parties" to make such a tool or package that does that - 
> if
> they want. you just made it possible to do that and have an example. :)
>
>> Does this have no problems?? Anyway, I hate damn patent :-(.
>
> welcome to the club. :)

Thank you for your advice.
Compiled out and pluggable which user can add/remove easily may be the
best solution.
But it have much to do :-&.
Patent issues related pinch gestures are not clear yet.  So it will
enter the todo list.  :p

-- 
BRs,
Kim.

------------------------------------------------------------------------------
Write once. Port to many.
Get the SDK and tools to simplify cross-platform app development. Create 
new or port existing apps to sell to consumers worldwide. Explore the 
Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join
http://p.sf.net/sfu/intel-appdev
_______________________________________________
enlightenment-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to