On Thu, 2006-08-10 at 15:32 +0300, Eero Tamminen wrote:
> > a. protects certain files from accidental deletion/ editing
> What is the problem of just by keeping the files with the genealogy datab=
> (sub)directory, so that they are both "protected" the same way, and if=20
> "certain" files need protecting, doing that with your desktop file manage=
> For protection you need all applications that can remove the files to
> participate in this protection scheme. You desktop file manager is the o=
> thing that at least needs to participate...
You're basically looking at it like I am. I'm thinking of a file
management app which creates a new user/group and makes them read only
to me as a user. That gets around my normal file 'browser' doing things.
So I can only manipulate them via this app, or as root. So I could
always give myself rw permission back again, but I couldn't accidentally
delete them. I could also copy, edit etc, but not with the original file
> > b. allows me to make text based comment files associated with
> > other files (to comment on PNG images for example)
> PNG files have comment section, but I don't know whether there are python
> bindings for manipulating them and anyway this differs from one media for=
> to another. Another solution would be to use e.g. one XML file per med=
> directory for this kind of information, like Digikam does.
As you rightly say, there is no standard implementation. Anyway it's the
digital equivalent of writing on the back of a photo - something to
avoid unless unavoidable (linguistic question: is that what's called an
> However, I'd like to know why the comment (notes) field in Gramps databas=
> for the media file is not enough, if the images (and therefore comments?)
> are genealogy related?
Well, plain text files have been around for about, oh, ages. Berkly
Database? I bet not as long. So from that we can assume that the same
will be true looking into the future. Much more important than
protecting my gramps database against future obsolescence is protecting
the information sources it's been built from.
All the information I have I am digitising with the lowest possible
technology, in most cases that's plain text, I'm considering using XHTML
zipped together with the referenced images - but then I might as well
use OpenDocument Format, and I haven't gone that way yet.
> > * There would be advantages in having a standard way that many
> > genealogy store their data, if it can be done for a branch as complex a=
> > Architecture, then we should be able to have one for genealogy. This
> > would:
> > a. allow different genealogy apps to guess where certain types of
> > files will be, images sources etc.
> > b. make transfering information from one genealogist to another
> > easier by being able to request a directory by name rather than
> > describing it.
> If the purpose of the media file comments is that they can then be used b=
> multiple genealogy programs, which programs you were considering?
> I mean, for this to be useful, those other genealogy programs would
> then need to support this scheme too! :-)
Not my problem really, but if the system was good some others might use
it. It should eventually have an xml schema and all that goodness of
I've attached an example of the schema i use now for all non-text files
I get. Of course it's taken me a while to get that far, and I'd guess
others have gone down similar paths and found similar solutions - so why
not make something standard so people don't have to continuously
> > c. assuming it would be a well designed structure, it would also
> > allow future users of the data we are collecting to more
> > easily understand the storage structure and the logic behind it.
> Instead of a structure, I think many people are nowadays frothing about=20
> tagging and automatic searching/indexing/catogorizing of their files
> (based on those tags, file system properties like date etc).
And i agree that it'll be great, but it's technology dependent. I want
lowest possible technology solutions. I'd rather rely on the likelyhood
that the words 'birth certificate' will be understood in 100 years time
than that a search technology can find the symantic relationship between
some scanned image somewhere in my file system and Auntie Dorris' birth
So there's some more of my thoughts.
Linux user: 372812 | GPG key ID: 21A8C63A | http://lithgow-schmidt.dk