I would like to add an addendum to http://bobklass.blogspot.com/2009/04/pdf-ifilters-in-sharepoint.html.
I was not aware of PDFLib's ifilter until I was near the end of my testing and procurement a few weeks back. Since I still had a pretty good test setup for this, and I needed to do some final QA, I decided to test PDFLib too.
Before you can test PDFLib you will need to get yourself a license from their sales group. They give you a reasonable testing period and the process is painless.
In today's test, PDFLib's installer didn't add both of the registry keys I mentioned in my last post, only the first.
I added the second manually. I would guess it is not important if you have MOSS.
My test set was not a very challenging bunch of documents (about 680 MB, 1200 documents). In some previous tests, Adobe was much slower (as anticipated) but in the ones today, they all three were similar. I previously was running the test after resetting the content. On today's I did full crawls (to save time??).
Interestingly, the quality of Adobe was much lower. The same search terms produced far fewer hits with Adobe while Foxit and PDFLit had the exact same results. It almost seemed anomalous, but my main objective today was to review the installation process (I have beaten that to death, no?) and give PDFLib a quick test, so I don't have time for further tests.
We already bought Foxit. I may have been able to save a little money with the way PDFLib is licensed, but I think I made the right choice because of Foxit's more foolproof installer. Hopefully I can add one more follow up article once I have been running Foxit in a large production environment for a while.
I was not aware of PDFLib's ifilter until I was near the end of my testing and procurement a few weeks back. Since I still had a pretty good test setup for this, and I needed to do some final QA, I decided to test PDFLib too.
Before you can test PDFLib you will need to get yourself a license from their sales group. They give you a reasonable testing period and the process is painless.
In today's test, PDFLib's installer didn't add both of the registry keys I mentioned in my last post, only the first.
- [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Office Server\12.0\Search\Setup\ContentIndexCommon\Filters\Extension\.pdf]
- [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Shared Tools\Web Server Extensions\12.0\Search\Setup\ContentIndexCommon\Filters\Extension\.pdf]
I added the second manually. I would guess it is not important if you have MOSS.
My test set was not a very challenging bunch of documents (about 680 MB, 1200 documents). In some previous tests, Adobe was much slower (as anticipated) but in the ones today, they all three were similar. I previously was running the test after resetting the content. On today's I did full crawls (to save time??).
Interestingly, the quality of Adobe was much lower. The same search terms produced far fewer hits with Adobe while Foxit and PDFLit had the exact same results. It almost seemed anomalous, but my main objective today was to review the installation process (I have beaten that to death, no?) and give PDFLib a quick test, so I don't have time for further tests.
We already bought Foxit. I may have been able to save a little money with the way PDFLib is licensed, but I think I made the right choice because of Foxit's more foolproof installer. Hopefully I can add one more follow up article once I have been running Foxit in a large production environment for a while.
No comments:
Post a Comment