Re: Updated WOFF2 spec is uploaded

I actually like what you just wrote: We don't know ahead of time what the
decoder will do. We know it is very likely to change something, therefore
remove DSIG because it is very unlikely to remain valid. I agree the spec
already tells you this but I think it could be more direct about it.

So ... maybe we could put something more like that in? (but in more
spec-ish language :D)

If the spec more directly pointed out the reasoning (improbability of DSIG
working + inability to predict decoder behavior ahead of time), I think I
would understand why DSIG MUST go away on first reading. Currently I think
I'd have to read it a few times if it was my first pass through the spec.

Ultimately it's fine as is; just trying to make it easier on the reader.

Cheers, Rod S

PS - just noticed a typo in: "between the *editros* draft used to prepare
the First Public Working Draft and the editors draft used to prepare"



On Mon, Apr 6, 2015 at 7:41 AM, Levantovsky, Vladimir <
Vladimir.Levantovsky@monotype.com> wrote:

>  Thank Rod.
>
> Not sure I completely understand your point here – what the spec is saying
> is that there can be many things changed as a result of the WOFF2
> encoding/decoding process besides glyf/loca conversion – table order is
> among them, which can be changed by either an encoder or a decoder, with
> the obvious effect on the table directory and table offsets. Not knowing
> ahead of time what the specific changes will be on the decoder side (but
> knowing that changes are very likely to happen) we are mandating that the
> DSIG to be removed by the encoder and indicate the modifying transform by
> setting the flag in the ‘head’ table.
>
>
>
> Again, we are mandating DSIG removal _*before*_ the modifying transform
> is applied [knowing that it will be applied], and the rest of the language
> in that paragraph is the justification for it – I am not sure what else can
> be said to make it more explicit. If you have any specific language to
> suggest, please do.
>
>
>
> Thank you,
>
> Vlad
>
>
>
>
>
> *From:* Roderick Sheeter [mailto:rsheeter@google.com]
> *Sent:* Monday, April 06, 2015 10:30 AM
> *To:* Levantovsky, Vladimir
> *Cc:* w3c-webfonts-wg (public-webfonts-wg@w3.org)
> *Subject:* Re: Updated WOFF2 spec is uploaded
>
>
>
> Small comment, it (diff
> <http://email.monotype.com/wf/click?upn=uA1u4oo8eyOFJ-2BcXH8sLkshuSA44ALP-2F0s9AodVupvecwkqKp048x1R2FhW-2Fo-2B79lRjfnyQjetY-2F6qxELXqai13-2FsAcexzPpt5vAAmwqvDm-2F-2FJP1hHOnR4IiNxq872N-2F_1G-2FjDAr7rgTstemRC6HwN-2F5kG9-2FqEkBGAZI-2BeK4lw-2BcbZ0p1LrE4rbHBf736K4l5m-2F0u8jeO8LzIgmTj-2BmqFwrbWWhVex6X5o-2B51V8dRVyHp-2BkpAVRAVdVO4raHC5UgwtDFJcORYhraFuPclQ7kI0vUNZyJfeAT475OHH9MI3ywCYfYfAlTBbIDWhhoDmnu2dtGsmOviYk33N2prRyzBKQ-3D-3D>)
> says several things *may* change and therefore we MUST remove DSIG. If I
> was reading it for the first time I think I would wonder why DSIG is
> removed when none of those things had changed.
>
>
>
> Presumably it is unlikely nothing would change and it keeps things simple
> to just kill DSIG; I wonder if we should be more explicit about this?
>
>
>
> Rod S.
>
>
>
> On Thu, Apr 2, 2015 at 12:29 PM, Levantovsky, Vladimir <
> Vladimir.Levantovsky@monotype.com> wrote:
>
> Folks,
>
>
>
> I have added the details about table reordering and how changes in table
> order and their corresponding offsets affect DSIG table (to justify its
> removal for CFF fonts), and fixed a couple of typos along the way.
>
> The new version is uploaded for your reviewing pleasure!
>
> http://dev.w3.org/webfonts/WOFF2/spec/
> <http://email.monotype.com/wf/click?upn=uA1u4oo8eyOFJ-2BcXH8sLkkZYEmDWlVONg6AVPNmy5M4Wk06JKOopu3DF5Gd6MVmc_1G-2FjDAr7rgTstemRC6HwN-2F5kG9-2FqEkBGAZI-2BeK4lw-2BcbZ0p1LrE4rbHBf736K4l5F0HnLCjtRWW8Pww6ff63WiC-2FA4gQ2g4jcTqkeiysmu8dxpyOcURR0pEkCxKdj8PJmzUiaJPe27Ia3qLGMLs9hpj7sSNMSz-2FD-2BuIUFLXZrV3HyTmYDTy96ca1F4w3vl6-2BwoRkmg9S0lhIAJeRMtkJdQ-3D-3D>
>
>
>
> Cheers,
>
> Vlad
>
>
>
>
>