Integrating tagging with workflow - batch?

System

Platform = Solaris 2.8
TeamSite 5.5.2 SP3, MetaTagger 3.5

I'm working on integrating MetaTagger into the workflow process for our 200+ intranet departmental sites and have come up with a set of questions / concerns / issues that I’m hoping I could generate some conversation around that may help me with the architecture / construction of this process. Feel free to jump in on any part:

Concern: performance --------
Our users are performance conscious. I'm concerned that if I put in a mandatory Tagging Review step into the existing workflow, there will be an overall negative impact based on the amount of time that it takes to construct the Set Metadata screen. Some of the users routinely just select a whole site (although it’s recommended against) and attach it to a job and, in this case, the resulting Set Metadata screen can take a long time to construct and can also be fairly intimidating to work with when lots of files are presented. Also, as far as I know, I can’t pare down the set of files presented in the GUI based on file type and I don’t want to show the metadata for .js, .css, etc. files.

Question: batch (iwmtbatch) processing with workflow. Good idea? / Bad idea? --------
Because of the issue mentioned above, I’m seriously considering how I can integrate MetaTagger’s batch processing into the workflow and make the review of the resulting processing an optional step. I’ve used iwmtbatch fairly extensively now with good results, so I think this is a good possible alternative. Here’s what I’d like to do: when the initial work on a job is marked ‘done’ by the author and the job is sent off for approvals, MetaTagger batch processing would be invoked on (only!) the modified files attached to the job, running as a background task that would have to complete before the files are submitted into STAGING, as the metadata needs to get into STAGING to be available to other workareas. I think I can use the mtbatch.update batch processing timestamp file to process only the modified files attached to the job…

Question: how to manage batch overwrite of user-modified metadata? -------
This is a particular touchy issue with keywords metadata. When a user reviews metadata and then modifies any of the MetaTagger-generated metadata, the subsequent batch update (or Regenerate via the GUI) of the file will overwrite any of the previous manual updates. The only solution I can come up for this issue is to have 2 metadata fields: MetaTagger Keywords and User Keywords and then merge them together when a full set of keyword information is necessary. Any other ideas?

Issue / feature request: batch processing of all the GUI-enabled fields.
I spent a bit of time constructing the GUI Set Metadata screen for setting metadata on our intranet content. 3 of the fields on the screen are hooked into MetaTagger processing and the other 6 fields are based on callouts built into the datacapture.cfg file. It’d be nice to have a command-line process that performed all of this processing in batch. As it is, when I call iwmtbatch on a file, it processes only the 3 MetaTagger based fields and then I have to script up a process for the other 6 fields. I don’t have any Templating experience beyond this Set Metadata screen. Is there something I’m missing (a batch process for running all of the processing associated with the Set Metadata screen), or am I going about this in the most logical manner available? Alternatives?


Thanks for any feedback!

Wally Box
Nike, Inc.

Migrateduser

My personal opinion: GIGO (garbage in, garbage out)

If you think that MetaTagger can be relied upon to give you the results that your organization needs without user intervention, then automate away!

Even if you have a manual, metadata review step, chances are that most users will simply hit "OK" unless they have already been convinced of the value of good metadata. You are tagging files for some reason, like navigation or search. (Otherwise, why bother?) The higher the quality of your metadata, the greater the value your content will have.

BTW, I think that MetaTagger does have a file filter option, but I am not positive.

Brinko Kobrin
Interwoven Staff Engineer