gbottu at ben.vub.ac.be (Guy Bottu) writes:
One solution to the problem would be to simply replace lookup with
something that uses getz directly, e.g. a perl script.
Even neater would be if there was a daemonized version of getz that
could take it's input from a socket, so that you could use the srs
indices on a remote machine.
Has anyone done something along these lines? I've started on the
lookup replacement and was thinking that I would write a perl wrapper
for getz that I could then run as a daemon on the SRS server (since we
have GCG and SRS on different machines), but if it already exists it
would be a bit stupid to reinvent it ...
Cheers /staffan
-----------------------------------------------------------------------
Staffan Bergh
Pharmacia & Upjohn, P14:5, SE-112 87 Stockholm, Sweden
email: Staffan.Bergh at eu.pnu.com + Don't let that horse eat that violin
phone: (int+46) 08 695 9884 + cried Chagall's mother
fax: (int+46) 08 695 4084 + but he kept right on painting
mobile: (int+46) 070 898 8829 + -- Lawrence Ferlinghetti
> Guy Bottu said:
> It is possible to use the same indexes for lookup and the native SRS,
> since GCG 8 lookup, GCG 9 lookup, SRS 4 and SRS 5 use the same format
> for indexes.
> There are however two problems :
> - the lookup of GCG uses the sdl files of SRS 4, while SRS 5 uses a new
> kind of files (i and is Icarus files). Although you need only one
> version of the indexes, you would have to keep SRS 4 and SRS 5 in
> parallel or decide to use only the older SRS 4. If you decide to install
> SRS 5, you will need to do a lot of hacking to make sure that library
> names, library numbers and srsfield shortnames are consistently the
> same in the sdl and Icarus files.
> - Reinhard Doelz found a way to hack the code of GCG 8 lookup to make it
> work with indexes generated by SRS 4 and recently I made GCG 9 lookup
> work with the indexes generated by SRS 5. One thing to do was to make
> sure that lookup reads only indexes for Sequence databanks and not
> others. Unfortunately, GCG has become touchy about its source code and
> we are not allowed to make the modified code public. I have no
> experience with the opposite operation, to make the native SRS work with
> indexes generated by GCG, but it should be possible. Note that you would
> have an SRS without sequencerelated databanks and without links. Note
> also that there is a good reason for using SRS 5 rather than SRS 4/GCG :
> it indexes a lot faster !
>> So, forget about it or take your start on a road full of obstacles...
>> Guy Bottu
