Ian Singleton

First Outline…

Posted in Uncategorized by iansingleton on February 9, 2010

What I envisage is a situation where I can see a tag cloud based on the information contained in the IPTC tags of my scans. To begin with, I classified all the information by two basic types – document type and subject. These can merge together in the tag cloud. From the first time I saw one, I’ve always thought that tag clouds are a brilliantly visual way of displaying information – not what the contents of the database might be, but also what their respective weightings are.

By selecting one of the value, I should then proceed to a second level of choice, this time effectively as a subset of the first tagcloud. If I chose the invoice tag, I’d expect to see a secondary tag cloud that looked like this one, a list of all the organisations from whom I’dreceived an invoice. Equally, if I’d selected a supplier first, then I might be presented with a list of options that include “invoice”, “letter” or “brochure.

From there, I’d like to see either a thumbnail of the relevant scans – like this – or a list based on my narrative summary of each scan. Of course, there’s going to be a problem when it comes to subcategories that have large populations, where there might be several screens of thumbnails. And while the thumbnail option is great for photos and visual images, it’s not going to be of great value with, say, the pages from a contract.

For text-based documents, it’s going to be important that the user is presented with a catalogue list of the documents which the user can scroll through to find the relevant data – or indeed, to be able to scrutinize even more closely with another search in order to extract the correct file.


Dunbar's Number

Posted in Uncategorized by iansingleton on February 6, 2010

is a theory by Robin Dunbar, about the number of friends you have in a social network. Any more than about 150 and it becomes unmanageable. More about that here. Although there is an interesting counter-argument here.

What about the number of backups you’re supposed to have?

Years ago, I had a graphics guy working at my house while I was in an office preparing a tender document. From the minute he called me about an hour after I left him to it, I knew that there was trouble.

He explained that as he’d stood up to leave the room, his chair had pulled the lead of a hard disc drive, which in turn had pulled the disc drive onto the floor. The drive was now making a strange noise and wouldn’t read. Was there anything on the drive, he asked, that was important?

My life, I answered.

Which was also backed up onto a different drive.

If the photo of my son going into the playhouse when he was four is priceless, what value do I put on the digital copy of the same photo which resides on a different drive? If there are five copies, is each of them worth a fifth of the original?

And how many copies am I supposed to make? What happens if the original and the backup are both destroyed. How many versions do I need to keep in the cloud so I can access them without recourse to the physical drives?

All of this material would fit onto any one of the SEVEN portable disc drives that I have. But at the moment, since all of these drives are sitting on my kitchen table, this isn’t a backup policy. It’ a burglary waiting to happen.

I can’t find a reasoned policy for how to distribute this information – I should give a number of these drives to friends and ask them to look after them for me.

There’s already a version of the compressed scans sitting on Zumodrive. And my end game for this material is for it all to be accessible from anywhere. Which raises a series of questions about passwords and access rights.

Trial And Error

Posted in Uncategorized by iansingleton on February 5, 2010

It’s been a frustrating week of (non) progress.

I’ve been aware for a while of IPTC tags – more here – but didn’t know how to strip the information from my Excel sheet and insert the relevant info into each of the JPEGs. Paul Beddoe has been invaluable in helping to overcome this problem. The first stab successfully wrote the contents of the Excel sheet – but wrote the metadata to the EXIF fields of the JPEG which aren’t readable by Photoshop or Flickr. Even though Paul’s hugely busy with his dayjob, he’s now managing to find some time in the evening to help overcome the problem with IPTC tags. If we can crack that, then we’re a long way towards having a workable application.