Maurice's blog

Some dancers (see http://perldancer.org) have expressed the wish to use more of Dancer's features in scripts accompanying webapps (in dancer-users@perldancer.org). To be a little more precise: they wanted to use the config information from a webapp outside the webapp.

The task

It's the first time I look into this.  I have some data in Excel and want it go into CCK fields. Concrete case: import pictures for the MIMO gallery and create the normal node and CCK fields easily. Of course I could save CSV or import to mysql table if I have to. 

Choosing the Module

First glance at 6.x modules indicates that Migrate seems to be most appropriate and ready for use. So, I start testing it on a test server.

Migrate

As with every rakudo star release I try to compile it on cygwin. So far, I could never do it. Today, I run into the same problems again as before. This time, i try to write down what happens.

Manifest test tells me that t/tools/install/testlib/phony.exe is missing. The file is actually mentioned in the MANIFEST, but it is missing in my unpacked rakudo star folder.

Beginners' Dance Class: http://www.youtube.com/watch?v=y64Ynur-veU

Perl turned 20 a couple of years ago. It has been with us since the Web has learned to crawl. Long before php, java beans and web application frameworks. There is tons of old perl out there, but there is also new perl code which runs on modern websites. Sometimes it is hard to find what is what.

I found a new toy: Dancer, the minimal effort perl web framework, see http://perldancer.org. Main advantage: easy to start even without any real webapp experience, good learning curve, nice community (still somewhat small) . Dancer code looks good. Some would even say elegant.

Anyways, I just noticed that I don't even need a session. I can just load info in class variable

our $memory_cache;

We talked about the decider today. The decider is supposed to select the media resources which shall be published. We agreed to use only one algorithm. By this I mean a relatively complex decision process and NOT a simple mapping rule, as for example:

if string "alpha" in field a then publish this resource to alpha.

Just a little, tiny note: When you run oaicat's demo from war file (version 1.42 or so), you might actually get no response for ListMetadataFormats if parameter is missing. I mean you get an error message that the parameter is missing, while the spec clearly states that the parameter is optional.

It seems that this issue has been fixed since current demo version at http://alcme.oclc.org/oaicat and also the upgraded version (available as jar and source) on my test box responds correctly.

Filesystem Structure

I guess I will go for a filesystem populated with many small xml files. Each file contains one record in native format and would be named according to the objId, e.g.

objId1234567.mpx
kueId1234.mpx
mulId12345.mpx

There should be only one record per objId, so probably I do not need several directories although this might get very big directory.

Updates

I guess I could write a little perl script to update the filesystem.

What happened before

So, I gave up installing tomcat on my laptop last week or so. Instead I use my ubuntu box at home now. That gave new headaches for a few evenings, but it was solvable and now it works reliably. At first I installed oaicat.war, then updated to 1.5.57 jar distribution and today, I was able to compile oaicat from source using netbeans. I guess I got my development system running. What is the next step?

Customaization Notes

I guess it is time to read the customization notes again. So where to look?

Today, Frank asked me for basic xslt and this week Andreas also showed more interest in XSLT than I expected, so I thought I write down some notes on how to get going with XSLT. I won't tell you what XSLT actually is, just info that you might be missing when you start out.

Another thing: One way of using XSLT is to convert xml to html or xhtml; i use it to transform xml data to xml data in another format.

Basic Info on XSLT