Re: [DotGNU]Architecture Neutral Distributed Format.

[Marco Manfredini]

Rhys Weatherley wrote:
> Hans-Olof Danielsson wrote:
>
>
>>Hi,
>>
>>It might be of interest to have a look at
>>
>>http://ddci.com/news_brief_ANDF.shtml
>>
>>describing the Architecture Neutral Distributed Format ( ANDF )
>>and compare it with Melody.
>>
>
> http://www.andf.org/ is the main Web site for this.
> Unfortunately, ANDF is a product of the Open Group
> (http://www.opengroup.org), and you have to pay money
> to get a copy of the specification.  Hence, comparing
> will be difficult.
>
> Does anyone have a link to the ANDF specification
> that does not involve "open" standards bodies that
> aren't actually open?
>

The TenDRA project included development of TDF, which is AFAIK more or
less ANDF.

I don't know of any TenDRA development anymore, but the data is still
available, including the TDF 3.x and 4.0 specs.

http://www.cse.unsw.edu.au/~patrykz/TenDRA/
http://members.nbci.com/tendra/

They have also a guide to the specification, which gives an overview.

Anyway, if undestand this right I find this strange. For example they 
want the translator not to make assumptions about the target machine, 
like about the number of bits in an long. But this means to me, that a C 
compiler should compile this in any case:

char x[sizeof(long)-3];

because he has no idea about the size of long. In C++ the situation is 
even more complicated, becase sizeof(long) could be used to select an 
explicit template specialization, resulting in completely different code 
for different wordsizes.

This could be done, but would lead to the situation that semantic errors 
could occur during the assembly of an application.

I will probably need to read that again.

Greetings, Marco