CCL:Re: James Stewart's article



 >> This copyrighted program is intended for users. It can be obtained
 >> from the QCPE.  There is no license fee for academic users, the
 >> license for commercial use is $2,000.
 >
 >The reason I never consider using the QCPE is that once I happened
 >to look at a list of prices for obtaining the software. It might be
 >that the hundreds of dollars these programs seem to cost is due only
 >to distribution fees, but the simple fact is that I can't afford it.
 >From my point of view, there is no QCPE. That's for rich folks.
 >
 >I understand that the QCPE could not afford itself if it did not
 >charge these fees, but from where I sit the net effect is the same.
 The above discussion about agencies such as  QCPE leads me
 to speculate about where we might be going in the future. QCPE
 represents a milestone in computational chemistry, since
 for over 35 years it has represented an organised and maintained site
 where users can obtain source code and documentation for
 what  I will call program "packages". About 15 years ago,
 we saw the start of the commercial chemistry software industry,
 but again the model was the same, ie "shrink wrapped"
 documented packages, although here the user often had to add their
 own code via custom  APIs rather than source code hacks.
 For perhaps 5-7 years, the  bio-informatics sector has run a
 different model of having many distribution centers based on
 the infra-structure of the Internet, and has not developed
 the same sort of commercial models, tending to rely much more
 on essentially free software (RasMol is an excellent example).
 Now,  I sense that the chemistry sector has the opportunity to
 move into stage 3 of this evolutionary process, which could be
 summed up by the description "chemistry objects". Whilst
 object oriented chemistry programming has been seen in
 many incarnations, the most recent advent of technologies such
 as  Java and most importantly  CORBA (common object
 request brokering architectures) and its Internet implementation
 suggests that this time it should succeed.
 I might add that most of the major chemistry software developers
 seem to be adopting the principle of chemical objects, although
 its is by no means clear that any of them will end up defining the
 SAME objects!  For example, there are about 15-20 molecule
 rendering "applets" written in Java, and yet there is no
 common object architecture between them at present.
 Which brings me to what role a "QCPE" might have in
 this new era. It might have to be a more pro-active one,
 in maintaining the chemical object namespace to go
 with CORBA (think of CORBA as an object equivalent
 of the DNS: domain name servers which we all rely on so
 much on the  Internet). It may also be much less of a
 "shrink-wrapping" agency and perhaps more of an API
 (application programming interface) organiser. Our
 own fledgling effort to start things up in this area
 is the OMF (http://www.ch.ic.ac.uk/omf/) but others
 are also travelling down this road. Lets hope we are
 all heading in more or less the same direction!
 Henry Rzepa. +44 171 594 5774 (Office) +44 594 5804 (Fax)