FW: [OpenType] maxSizeOfInstructions

[Levantovsky, Vladimir]

Forwarding Jelle's response on the OpenType list:

-----Original Message-----
From: listmaster at indx.co.uk [mailto:listmaster at indx.co.uk] On Behalf Of Jelle Bosma 
Sent: Sunday, June 21, 2015 2:06 AM
To: listmaster at indx.co.uk
Subject: Re: [OpenType] maxSizeOfInstructions

Message from OpenType list:



> On 19 jun. 2015, at 15:35, Levantovsky, Vladimir <Vladimir.Levantovsky at monotype.com> wrote:
> 
> ISO spec is just about to undergo another revision and the amendment will be opened by the end of next week. If you can submit a proposed 'maxp' change I can add this to other amended items and considering we have no objections to the proposed changes we can make them into the spec fairly quickly. Please send the proposed spec update to the mpeg-OTspec at yahoogroups.com list.

Hi,

I very much object to changing the specification of the maxp. There is a specification which does not have the slightest ambiguity about how the maxSizeOfInstructions should be calculated.

The TrueType specification version 1.0 of 7 July 1990 says: byte count for glyph instructions.
The current OpenType specification: Maximum byte count for glyph instructions

The note to include Font and CVT programs (fpm and prep) for the maxStackElements, has been in all specifications of the past 25 years too. 

And it make sense: why would you cache the size of the instructions in the prep or the fpgm if that equals the size of the tables anyway. How would a rasteriser not know the size of the instructions when it fetches these tables to execute the instructions? That is not true for the maximum stack. And that is why the maximum stack includes the stack of the prep and fpgm. 

Why would we now be expected to update our fonts, production and test tools, for a value that is evidently not used by Apple and Microsoft, and which, if it is calculated how the published specs have told us to calculate it for almost 25 years, works well in all known applications.

Best regards,
  Jelle Bosma




List archive: http://www.indx.co.uk/biglistarchive/

subscribe: opentype-subscribe at indx.co.uk
unsubscribe: opentype-unsubscribe at indx.co.uk
messages: opentype-list at indx.co.uk