Re: [greenstone-users]

From Jack Park
DateWed, 28 May 2003 07:41:14 -0700
Subject Re: [greenstone-users]
In-Reply-To (20030528051612-GA22108-wesson-cs-waikato-ac-nz)

This is REALLY interesting! I shall now revisit using greenstone on wintel.
Cannot say why it happened, but greenstone (or NT) hauled off and just
started overwriting its old directories, trashing everything in site. I
don't recall the numbers (this occurred 2 years back) but it went from
something like, say, 16,000 back to around 50 in one fell swoop! And, I had
some really cool stuff archived.


At 10:16 PM 5/27/2003, John R. McPherson wrote:
>On Tue, May 27, 2003 at 10:01:35PM -0700, Jack Park wrote:
> > Deepak,
> > I ask because, running on Windows NT, I managed to "overflow" something to
> > do with windows: too many directories. That is because greenstone, running
> > on unix boxes, compresses the directories into one object, but on windows,
> > it does not, and eventually, the archive does something that windows
> > doesn't like.
>because greenstone still supports windows 3.1 (FAT16 disk partitions),
>filenames and directories are limited to 8 characters (plus 3 letter
>extension), which is partly why lots of directories get created.
>Maybe FAT32 has problems with lots of files/subdirectories but I
>would have thought windows NT (NTFS) would be ok.
>If you have a collection that does have lots of documents, you can
> add:
> groupsize 200
>or similar to your collect.cfg. This means greenstone will put 200
>import documents into each archive .xml file. For example, we do this
>with email collections, where there are lots of relatively small
>documents. This can help reduce the number of directories and
>subdirectories required to store the archives.
>John McPherson

XML Topic Maps: Creating and Using Topic Maps for the Web.
Addison-Wesley. Jack Park, Editor. Sam Hunting, Technical Editor

Build smarter kids, not smarter bombs.