Re: OSHIP & Python 2.5
On Mon, 2008-08-25 at 22:33 +0200, Roger Erens wrote: > In the current installation for linux page > (http://www.openehr.org/wiki/display/dev/OSHIP+Installation+-+Linux ) > I > have the following remarks: > > 1) if you advise to create a directory called .buildout, it would be > more consistent to have as eggs-directory > /home/tim/.buildout/buildout-eggs instead of /home/tim/buildout-eggs. Hmmm, not sure what you mean here. My suggestion (taken from the buildout and zopeproject instructions) was to create a directory called .buildout in your home directory with the file default.cfg in order to tell buildout where to store eggs that may be reused across many projects. Of course I did say to put: [buildout] eggs-directory=/home/tim/buildout-eggs inside default.cfg (per above instructions) since this would seem to be consistent with other buildout based projects. Using the '.' leading directory names is pretty much a standard for configuration directories/files. So you could easily put: [buildout] eggs-directory=/home/tim/.buildout/buildout-eggs in default.cfg but then .buildout is no longer just a configuration directory. Your choice though. > > 2) I think there is no need for the step '$mkdir sandbox' as this > directory will be created by the next step if not present: > '$virtualenv > sandbox' > Thanks. Actually I didn't realize that virtualenv would create the directory if it didn't exist. > 3) also, the step '$mkdir oship' can be left out if the current svn > checkout command is concatenated with ' oship' (mind the space) > True. > 4) maybe it's interesting to note that executing 'oship-ctl fg' from > within the ~/sandbox/oship/bin directory fails because of inability > to > find zdaemon.conf. The server should be started from within the > ~/sandbox/oship directory. Also true. That is why the instructions leave you in the oship directory and say to execute: $bin/oship-ctl fg If you think an explanation is needed then it can be added. > > 5) finally, I wouldn't use the name 'sandbox' for the virtual > environment. Developers will probably have multiple virtual > environments > created, and 'sandbox' is not very specific. I'd choose something > like > 'oshipenv'. > True to some extent. But it is unlikely that developers will have more than one virtual environment for each version of Python that they wish to work with. For example you may want a virtual environment for Python2.4,Python2.5,Python2.6 and Python3. But each project based on each version of Python would be included under that virtualenv. Something like: sandbox2.5 - oship proj1 proj2 sandbox2.6 - oship proj1 etc. This lets you reuse consistent environments. > In the Basic Usage Guide > http://www.openehr.org/wiki/download/attachments/2949253/OSHIPUsage.pdf?version=5 > > 6) the first note states that 10 sample ADL files should be located > in > the oship/src/oship/import_adl directory. I only found one: > openEHR-EHR-EVALUATION.goal.v1.adl > VERY true. The usage guide hasn't been updated at all since the LNCC Workshop. There we decided to focus on one archetype that has more functionality instead of 10 archetypes with no real functionality. > I didn't get further in the Guide/Tour. We should probably delete the usage guide and combine the Linux and Windows install guides into one document that would include the basic parts of the usage guide. Also, the documents that are currently in the oship/src/oship/docs directory are completely out of date. The management of the installation text is completely out of sync. I would like to say that the document oship/src/oship/docs/Installation.txt is the 'source' information and then I could easily copy/paste it to the wiki page. Roger, since you have SVN access would you please update the Installation.txt file along these lines? Alessandro currently has a problem with Windows installation that we have not tracked down yet so the Installation.txt file should say something to the effect that Windows install instructions are coming soon. I can then do the copy/paste to the wiki and the wiki will always be secondary to the Installation.txt file (as it should be). Thanks, Tim PS. I still have a lot going on right now but as the week progresses I will get more info on the results of the LNCC workshop out to all of you as well as some planning. I can tell you that things are progressing but I have little time to concentrate on writing about them over the next day or so. -- ****************************************************************************** Timothy W. Cook, MSc Health Informatics Visiting Researcher, Clinical Decision Support Systems Fluminense Federal University, NiterÃi,RJ Brazil Honorary Research Assistant, CHIME, University College London, London, England, UK ******************************************************************************
Attachment:
signature.asc
Description: This is a digitally signed message part
_______________________________________________ ref_impl_python mailing list ref_impl_python@openehr.org http://lists.chime.ucl.ac.uk/mailman/listinfo/ref_impl_python