Maximilian,

I think that there are many advantages of things (in general) being formally 
defined. Especially when it comes to documents and processing
standardizing formats using XML can be essential. If you have ever tried to 
convert different  formats between applications or implemented parsers for 
whatever 
documents users might come to you, you will experience what I mean.
It will be of great help, if you can at least tell if the document is 
syntactically valid or not. Actually, in the 'real-world' of 
computing/communication standardization is a must and it 
has been there since ages. Look at the RFCs [1] for an example of the benefits 
formalizing things, or what would you say if your mail program did not support 
SMTP and IMAP correctly?

I think the fact that you dropped SOAP and that not many people are using 
workflow tools might arise from the same frustration people might experience 
when they try to use these protocols. Then they give up on the matter, and 
maybe nobody tries to even report the problems.


Best
Michael


[1] http://en.wikipedia.org/wiki/Request_for_Comments

Am Mar 22, 2010 um 2:56 PM schrieb Maximilian Haussler:

> You can send the REST queries via HTTP POST, right? Then there is no limit 
> anymore.
> 
> It is nice if something is formally defined but I don't see why this by 
> itself constitutes any advantage for practical use (okok, no, I also don't 
> want to get into a SOAP vs. REST discussion, I have used SOAP once and then 
> dropped it forever and will definitely never use it again).
> 
> There are workflow tools in bioinformatics but the very large majority of 
> developers don't use them currently.
> 
> cheers
> Max
> 
> On Mon, Mar 22, 2010 at 12:09 PM, Richard Holland <[email protected]> 
> wrote:
> REST is only suitable for queries with a small number of filters or 
> attributes that can fit within the maximum length of an HTTP URL. It is also 
> unsuitable where large numbers of values need to be specified in a filter, 
> e.g. by uploading a file of IDs. SOAP solves both these problems and has the 
> added bonus of being formally defined therefore much easier to generate 
> automatic interfaces to - that is, if it validates! :)
> 
> cheers,
> Richard
> 
> On 22 Mar 2010, at 11:05, Joachim Baran wrote:
> 
> > Hi!
> >
> > On 22 March 2010 10:33, Michael Dondrup <[email protected]> wrote:
> >> I would like to second Pierre in asking for improving the interoperability 
> >> of the BioMart SOAP interface.
> > Is it not simpler to use the RESTful interface? I mean, you basically
> > send just a single string to the mart that describes the filters and
> > attributes of your mart-query, and you directly get the result back.
> >
> > BW,
> > Joachim
> >
> > --
> > B.1079 Michael Smith Building
> > Faculty of Life Sciences
> > The University of Manchester
> > Oxford Road
> > Manchester
> > M13 9PT
> > United Kingdom
> 
> --
> Richard Holland, BSc MBCS
> Operations and Delivery Director, Eagle Genomics Ltd
> T: +44 (0)1223 654481 ext 3 | E: [email protected]
> http://www.eaglegenomics.com/
> 
> 

Reply via email to