Should we actually use the nlm code as the "base"?

[Demian Hess]

The NLM version of the code (so called data dictionary) is a pretty good code-base (aside from my typical goofy packaging). i like the content model better--not element centric--and it even comes with a DTD!

also, it captures things like filename and line number

and the app is a nicer experience--applies a transform and doesn't require the use of a catalog if you don;t have one.

minus is that it dropped support for annotations--could we work to add annotations into the data dictionary app rather than try to port features of that code base into dtd analyzer?

[Demian Hess]

and to be clear--call the nlm version of the code the "dtdanalyzer" and change the packaging to make it make more sense...

--
You received this message because you are subscribed to the Google Groups "DtdAnalyzer" group.
To view this discussion on the web visit https://groups.google.com/d/msg/dtdanalyzer/-/9hKVDyRX34sJ.
To post to this group, send email to dtdan...@googlegroups.com.
To unsubscribe from this group, send email to dtdanalyzer...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/dtdanalyzer?hl=en.

--
Demian Hess

Avalon Consulting, LLC
527 Maple Avenue East, Suite 200, Vienna, VA 22180

Mobile: 301-943-8307
Fax: 845-367-5496
he...@avalonconsult.com

[Chris Maloney]

Hi, Demian,

More and more confused -- sorry!  Let me make sure I know which version is which ...

I have two versions:  

1.  dtdanalyzer - this is what is on Github right now at https://github.com/Klortho/DtdAnalyzer.  Namespace is gov.ncbi.pmc.dtdanalyzer.

2.  datadictionary - this is what you sent recently, and I just replied on a different thread with an attachment that is basically the same bundle,   Namespace is gov.pubmedcentral...

So, when you say, "the nlm version", which are you referring to?  I guess, after re-reading your emails a couple of times, it's clear you mean the datadictionary; but I just want to be sure.

If so, it sounds like it might make sense, but I'll have to take a look and make sure we can justify it.

Cheers!

[Demian Hess]

i wrote the 'data dictionary' for nlm. always hated that name but it's what the qoman rhere (i forget her name) kept saying she needed. 

Sent from my iPod

[Chris Maloney]

Hi, Demian,

I had a look at your datadictionary version yesterday, and I agree that it's got a lot of nice enhancements in terms of the output that's generated.  The output format isn't much different, and I don't think the few differences there are matters very much.  If anyone wanted the old output format, it would be trivial to write a transform to "down-convert".  The one problem that's not minor, though, as you mentioned, is the lack of support for annotations.  I wonder how much trouble it would be to add that back in.  Hopefully it would just be a matter of integrating that portion of the DtdAnalyzer code back in, and it wouldn't be too painful.  I will take a look when I get a chance.

This looks very nice, I do like the filenames/linenumbers and especially the fact that you capture entity definitions.

Chris

[Demian Hess]

I think it'd be easy to put the annotations in--just implement the callbck for comments to do the parsing and stick the info somewhere in the model. would be nice to capture annotations for element definitions that haven't been seen and then, when the definition is found, add the annoations to that model for output. 

I'm woefully behind in everything--poke me with a stick when i'm not being responsive...