> In article <3gv4m7$17m at mother.usf.edu>, gquinn <gquinn at com1.med.usf.edu> writes:
> > I'm posting this as a follow up to my previous posting, which
> > enquired whether it might be possible to compile the GCG source
> > code on Solaris X86, which is a UNIX operating system that runs
> > on a PC. After receiving several replies on the subject from people
> > more knowledgable than myself, I can say that the answer is
> > (probably!) yes. Both the C and FORTRAN files are likely to compile
> > 'straight out of the box'. I think that this is not a trivial
> > point, since X86 is currently selling like the proverbial 'hot cakes'
> > since SUN made it very affordable. Additionally, the SUN FORTRAN
> > compiler is about to be added to the list of 'cut price' software
> > they're knocking out, so the cost involved in setting up a PC for
> > SUN UNIX and compiling the GCG program would be effectively limited to
> > the cost of the program. This UNIX box could also act as a server.
>> Hsm... sounds like a good way to violate just about every section of the
> license you signed with GCG when you bought the package! Of course GCG might
> be happy to renegotiate the contract for you to do this!
>> with apologies for raining hard on your parade and all that.....
>> jasper
I wasn't going to bother to reply to any more postings on this subject
'cos it is rather trivial, but seeing the other crap that's being posted
in this group (Law Services, Valentine confection etc) at least this
relates to GCG.
Jasper, I think that you need to check your licensing documentation for
GCG! My understanding is that I can buy GCG and put it on any single platform I
care to, though binaries and support are only supplied for selected
platforms (there have been several inquiries to the newsgroups from
people compiling GCG for other UNIX flavors, eg. SUN BSD).
The question of whether it's on MULTIPLE machines at one site is
another question altogether, and in my original posting on this
subject I suggested that this area would be something of a minefield.
To my knowledge, GCG is neither a charitable institution nor a shareware
company, and copies must be licensed, if for no other reason than
that further software development requires it. You appear to be
assuming that because the program would be running on a PC that it
would be unlicensed!?! Why?
The purpose of my post was to point out that an expensive machine is not
required to run GCG (NOT everyone can afford one) and that someone
who does not have good (or any) network connectivity would likely be
able to compile the GCG code 'out-of-the-box' on a platform that would
require minimal cost to set up, and allow the operation of a local GUI.
I think that the whole thing IS 'doable'.
Whether this would have further application for general standalone usage
would remain to be seen. Several people have pointed out the 'disk access'
problems associated with this (NFS mount traffic, disk cost, etc)
and on reflection this may be very real. However, since the binaries
are only about 60-80meg (very approx), the hard disk space needed would
not be gigantic unless databases are put on. The online BLAST server
in V8.0 works great for me and obviates the absolute need for a local
database (actually Email has done this for a few years). Unless you need
to do a FASTA scan (and you can do this by Email too), I am finding it
difficult to understand the logic of having a local database copy,
taking up more than a gig of expensive disk space, that needs to be
updated periodically. I can get a retrieve command reply from NCBI server
in often less than 10 seconds. (WPI won't even 'fetch' sequences except via
the awful 'seqed').
Probably 'bout time to kill this one, chaps.
