Agreed.

My point was that on embedded devices without floating point and with minimal memory an implementation
that tries to be compliant will have a tough time because:
  • can't implement floating point because no hardware
  • can't do a good job at s/w emulation for floating point and decimal due to memory constraints
For many apps, integer numerics can suffice. But rather than create a non-compliant extension, I was proposing that the spec allow such configurations by providing an integer numeric selection pragma.

I hope this clarifies things.

We too are implementing in parallel with the ES4 spec development and sometimes it is painful as we have to rework our code to track the spec as it changes. Other times, I hope it is helpful as we get real-world, hands-on testing of proposed language features before they are set in concrete.

But then again -- we're not an industry Gorilla and no one will be too worried if we ship something early ;-)

Michael

Brendan Eich wrote:
On Dec 13, 2007, at 3:09 PM, Nathan de Vries wrote:

  
On Thu, 2007-12-13 at 08:07 -0800, Michael O'Brien wrote:
    
I believe this is an important issue that will force non-standard
implementations...
      
On the same topic, it seems that Adobe are planning on introducing the
decimal type into the Flash Player before the ES4 spec. is finalised.
They're asking Flash developers at the moment what their requirements
are for "decimal math".

Is the spec. at the point where implementers can start preempting it?
    

Not quite, but no one is going to preempt -- the early implementors  
will have to shift if they want to claim conformance to the final  
spec. We've done this in the past in Mozilla and it has been tolerable.

Michael's point seemed to me to be more about divergent long/ulong  
semantics and no spec, not early implementations trumped by the one  
true spec.

/be

_______________________________________________
Es4-discuss mailing list
[email protected]
https://mail.mozilla.org/listinfo/es4-discuss

  
_______________________________________________
Es4-discuss mailing list
[email protected]
https://mail.mozilla.org/listinfo/es4-discuss

Reply via email to