Re: medical image database

From: Philip Lijnzaad <lijnzaad_at_ebi.ac.uk>
Date: 21 Aug 2000 14:37:34 +0100
Message-ID: <u71yzizqe9.fsf_at_o2-3.ebi.ac.uk>


> Second issue you need to take care is the application security. In
> reality, you would need to orchestrate your entire application security
> around the image field in your database - like who can edit it, read it,
> dump it, etc. So design your application security schema with
> considerable thought.

yes, absolutely. But what do you mean by ''orchestrate your ... security around the image field '' ? Security applies to all the data, not just the images. And I suspect the security of the person records is in face even more important than that of the image data.

> On the whole, storing the images as files in the file system is dumb.

no, I disagree. There are good reasons to keep these images in plain filesystem files. The most important reason being the interoperability with lots of (home-grown, legacy) analysis programs, which general will rely on the names of files in order to access them. I know of no database where you can say UPDATE test_scans apply Gaussian(IMAGE, 1,2,3,4) WHERE capture_date BETWEEN 1-May-1998 and 14-may-1998, or something of the sort. By storing everything in a database, you add an additional nightmare of exporting/importing images from/into the database, with no additional benefit.

Lastly, images need not be transactional, as long as you have backups. (but ideally, the filenames should be transactional ...).

I'm sure there are specialized DBMSs for image mangement, but I'm not an expert on that.

> "I
> Love You" virus ate a lot of such image banks all over the world, a few
> months ago. Encapsulating them in some repository would have prevented
> this.

And so would choosing a decent operating system and/or decent backup mechanism in the first place! Sorry, but I find this a totally ridiculous argument to put images inside a database. Especially given the security considerations already mentioned, and the very weak track record of the software vendor that has exposed quite a few people to viruses of the IloveYou type.

                                                                      Philip

-- 
When C++ is your hammer, everything looks like a thumb. (Steven Haflich)
-----------------------------------------------------------------------------
Philip Lijnzaad, lijnzaad_at_ebi.ac.uk \ European Bioinformatics Institute,rm A2-24
+44 (0)1223 49 4639                 / Wellcome Trust Genome Campus, Hinxton
+44 (0)1223 49 4468 (fax)           \ Cambridgeshire CB10 1SD,  GREAT BRITAIN
PGP fingerprint: E1 03 BF 80 94 61 B6 FC  50 3D 1F 64 40 75 FB 53
Received on Mon Aug 21 2000 - 15:37:34 CEST

Original text of this message