[Mycroft] 2 kinds of questions

[Alan Bramley]

Hi Matt.

For the first question I think I can help. We have already given a 
little thought to plugins for subscription-only search and decided it 
would be better not to include them. The reasons being...
1. We can't easily make the limitations clear and 'unknowing' 
downloaders could easily assume the plugin was broken.
2. In order to write, test and maintain these plugins we would need to 
take subscriptions ourselves.

For your second question I think my colleagues on the team may probably 
give you more helpful information...

However, if I understand this correctly you are needing an input field 
for each plugin you select so that you can form the search terms 
individually  to match boolean logic of each target engine. The 
immediate question is how many search engines will the user select? 
Obviously this is unknown, so the software would have to auto-generate 
an new input field as each new plugin was selected. The next problem is 
what happens if somebody selects a very large number of plugins? Perhaps 
it would just have to be fixed at (eg.) 5 max. It's a valid request but 
it looks to me like it would be tricky to implement within the existing 
Mycroft. So, maybe it needs to be a new feature, a sort of "Mycroft 
Advanced".

Regards,
Alan


Matt Price wrote:
> Hello everyone,
> 
> I have 2 sorts of questions about mycroft plugins, both about
> bibliographic search engines....
> 
> 1) ever since I discovered mycroft I've been enthusiastic about adding
> library searhch engines.  Yesterday I wrote a couple of little .src
> files for subscription-only  databases -- the online Oxford English
> Dictionary, and the enormous JSTOR archive of scanned-in journal
> articles from the humanities.  I think it should also be possible to
> write a plugin for the ISI Web of Science... 
> 
> I was wondering whether there's any point in submitting these plugins
> to you guys.  The JSTOR plugin in particular involved a fair bit of
> hunting around in the results pages in order to figure out how to
> construct the input fields; so I'd like to save other people the work
> of piecing together the logic of their search engine.  But my plugin
> will only work for people working on computers that jstor recognizes
> as belonging to institutions tht have subscriptions to the archive...
> so it's bound to disappoint some of the folks who try to use it.
> What do you think?
> 
> 2) I also finally managed to install the SearchSideBar on firebird
> yesterday (needed root privileges, duuh), and I really like it.  But
> one thing I notice is that the complex bibliographic search engines
> have non-identical boolean syntax -- so it's really hard to combine
> searches in, say, the University of California Library, the Library of
> Congress, and the Harvard library system.  If it were possible to have
> more than one search field -- say, Author, Title, and Subject/Keyword
> -- then combining searches would be much easier.  Writing the search
> plugins wouldn't be that much harder then it is right now, provided
> the Sherlock syntax is flexible enough to allow for multiple user
> inputs.  (Since it's clearly modelled on the HTML "FORM" syntax, I
> imagine this should be possible).  But modifying the code for the
> SideBar might be much more complicated.  Do you have any idea 
> how
> feasible it would be to write a BibSearchSideBar based directly on the
> SearchSideBar code, but adding a couple of extra input fields & maybe
> changing the location of the search plugins?
> 
> Anyway, thanks for taking the time out to red my mails!  
> 
> matt
> 
> 
> _______________________________________________
> Mycroft mailing list
> Mycroft at mozdev.org
> http://mozdev.org/mailman/listinfo/mycroft
>