micha at amber.biophys.uni-duesseldorf.de micha at amber.biophys.uni-duesseldorf.de
Sun Sep 18 15:06:51 EST 1994

Reinhard Doelz (doelz at comp.bioz.unibas.ch) wrote:
: Przemko (przemko at reks.uia.ac.be) wrote:
: : When I use Mosaic to access various databases to do searches, 
: : sequence comparisons etc. I discovered a problem. It seems that I cannot 
: : submit actual sequences. See, I am not going to retype 2000 bases, or to
: : submit a Fasta formatted multiple alignement file for a 
: : domain search, Sequence Logos etc.
: : Any idea how I could place a file in the search quesry of Mosaic?
: : Even cut/paste does not work if a query is too large...
: : HELP!

: Short answer: No way, cut+paste is the only method. 

And even in this case I am not sure if 2000 bases, fasta-formatted of course,
will make it to the other side (does anyone know what is the limit on the
number of elements=lines of a textarea field?).

: Long answer:  

: These 'queries' have 
: to be in a given format, and this is a problem for file transmissions; 
: as the format definition (computer colleagues know this as MIME definition)
: is cruical for success. Obviously, the different formats applied to 
: biological sequence data would need to be known to Mosaic as the client.
: Other distinct areas or disciplines will need other formats, and, in the
: end the general-purpose nature of http would be endangered. 

Reinhard may correct me if I'm wrong, but I think form requests are sent
as "text/html" MIME type, and so the only chances are to use plain sequence
without comments or annotation data for compatibility - or to write a Bio
WWW client, using the NCBI sequence data specs internally (but this is
off-topic here; jumping to bionet.software.www)

[ explanation of http protocol deleted ]

Another question to Reinhard:

: We are currently developing the HASSLE receptors and -reflectors for the 
: http protocol, in order to get our HASSLE system integrated into WWW. 
: It might take a while, though, as software still has to be finalized. 

You mentioned the synchronous nature of http above, in contrast to the
asynchronous "background" nature of the provider's computations that
is implemented in HASSLE - how do you solve this, reconnecting to a 
WWW client later won't work, and authentication of WWW clients is
quite another issue? That's not a mock-question, I had similar problems
and only saw the chance to take to the good old e-mail for return, so
I gave it up!

: Regards
: Reinhard 


	Michael Schmitz

More information about the Bio-soft mailing list

Send comments to us at biosci-help [At] net.bio.net