On 25/03/2015 22:36, Joseph Wright wrote:
> On 25/03/2015 21:52, Karl Berry wrote:
>> Sigh.  Again: my point is that one way is verifiable and debuggable, the
>> other is not.  If all that's stated is "7.0.0", and a user's result is
>> different, there is no way for the user to know if it's because Unicode
>> released two different files under that version number, there was an
>> error in the download, something changed or went wrong in the LaTeX
>> processing code, or ... who knows what.
>>
>> With a factual, verifiable, piece of information about the input files
>> used, any problem can be easily diagnosed, instead of having to be
>> guessed at.
>>
>> (Not that it's likely to matter in practice, I grant you.)
> 
> This is done using pdfTeX's \pdfmdfivesum. I won't get a snapshot done
> today but will in the morning: check DropBox in about 12 hours!

I note that both of the files with versions also have dates. I wonder if
these might be as (or more) useful to most end users than MD5 sums.
After all, a version/date combination is what we use for LaTeX packages,
etc.
--
Joseph Wright