Re: new proposal to vector intrinsics

[Tobias Burnus <burnus at net-b dot de>]

Sa Liu wrote:
> > Given this interaction, I think that we need to see a design for the
> > implementation of parameterized derived types, rather than sending you
> > off on a mission to return with the job completed.
> >     
> Due to  limited Fortran skills in our team, could you please supply a list of requirements concerning PDT? I was just aware that PDT should be involved in OOP in Fortran.
>   
As starter for reading about these I would suggest:

"The New Features of Fortran 2003" by John Reid,
ftp://ftp.nag.co.uk/sc22wg5/N1601-N1650/N1648.pdf
Section 2. (2.2, 2.9 and 2.10 can be ignored.)

I also would recommend the 440p-thick book "Fortran 95/2003 Explained" 
by M. Metcalf, J. Reid and M. Cohen, Oxford, 2004;
http://www.oup.com/uk/catalogue/?ci=9780198526933 (41.83 EUR at 
www.libri.de)
I think it nicely summarizes Fortran 95 and Fortran 2003 features 
including OOP and PDT; I find it very readable and a good language 
reference. It is concise and useful especially if one can already 
program; it contains also nice examples and is not too expensive.

[Another good, 720p-thick book would be "The Fortran 2003 Handbook" by 
Adams, Brainerd, Hendrickson, Maine, Martin and Smith, Springer, 2008, 
which has not yet been released but is scheduled for June 2008. (Authors 
include the Fortran 2003 editor and other standardization members, 
testsuite writers etc.).]

Before implementation one should also check whether something relevant 
has been added/changed in the draft of Fortran 2008 standard regarding 
OOP and derived types, but I don't think so. I will do so after Paul 
posted the first draft in the wiki.

Tobias

PS: Fortran standard (or its drafts) are available via 
http://gcc.gnu.org/wiki/GFortranStandards