Re: [VM] Info node "External Messages"

[Uday Reddy]

address@hidden writes:

>  > So,
>  > "evidence" is practically useless!
> 
> this is not a universal conclusion. If it were, then reverse engineering
> closed source software would be impossible, because there you do exaxtly that:
> You try to look for evidence (either by sniffing, disassembling, etc.) that
> the software does a certain thing.

We are going off-topic, so I won't belabour this.

Reverse-engineering works only as long as the original developers stick to a
particular behaviour.  If the behaviour is not documented and published,
essentially being made into a contract, the developers can change the
behaviour at any time and break interoperability.  That was the foundation
of the celebrated EU vs. Microsoft case.  Even though Microsoft made its
source code available, EU said that it wasn't enough.  The specifications
needed to be published.  (One wonders whether Microsoft had any
specifications at all in order to divulge them.  They learnt the hard way
that they had to write specifications, if only to help their own
competition!)

Coming to the issue at hand, I doubt if Thunderbird developers are any
better at writing specifications than the Microsoft developers.  The best
you will get is stuff like this https://developer.mozilla.org/En/Folders.
The reason we are able to inter-operate with Thunderbird is not that
Thunderbird is well-documented, but rather that they have chosen to base
their folder format on a standard RFC, the same one we use. (RFC 4155).
Nothing stops Thunderbird from deviating from the RFC, but in all
likelihood, they won't.

On the other hand, for IMAP cache folders, the very first line is a
violation of the RFC.  So, in a way, they are warning us that this is not an
RFC-client folder and we shouldn't mess with it.

Cheers,
Uday



> 
> Even for well documented software the conclusion only applies if what
> you want to to with it is part of what is documented, which is not always
> the case.
> 
> Granted, in the context of this discussion (viewmail/thunderbird) the 
> conclusion
> does hold but only because both are well documented and what he wants to do
> is covered in it.
> 
> Bye,
> 
>    Moritz
> 
>