Sinzui

It should be known that in the grand schema of all things Storage, I am the toady to s/master/maintainer/ Seth. Sutra, a meta data database, is my concept to solve the limitations the Medusa faces, and address the storage concerns that Marco brought to my attention regarding bookmarks and browser history. I intend to get Medusa working well as a file indexer/search tool for posix file data. I will replace Medusa parts with Sutra parts to provide richer meta data and search services. Seth will likely replace Sutra parts with Storage parts to provide better content management and search features.

The W3 published a slew or RDF drafts today. Lots to read. I've been reading more about strategies for mapping RDF to RDBMS. Reiteration appears to be too complex to solve. I'm not going to pursue it since better minds than mine have failed. Still, I've got a good idea how to make subject, predicate, object graphs work in a DB using the ideas of RDF Schema and OWL.

Interestingly, National Public Radio (NPR in the USA) had an article on magnetic resonance imaging (MRI) the brain as people solve problems. The network of cells that knows tools is different from the network of cells that knows animals. Thinking about doing something (motion) often leads to the tool network. There is a high cost to get the name of a tool the first time. Each subsequent attempt to recall takes less energy, with evidence that the network is changing its connections to get the answer quickly.

/me thinks each cell keep state of last connections used.

/me thinks each node changes the sequence of index of connection so that the most recent used link is a the top of the stack. subsequent calls are faster because there are fewer steps in locating matching nodes.

/me thinks helper cell must test if four or more neuron connections always occur in a sequence, a connection is made from start to finish to shorten the path.

/me thinks timers and triggers restructure the graph to optimize paths based on proven sequences of connections. Nodes provide feedback to grade the response.

I recall that the brain storing images of objects as images, a circle in a real world matches the circle of neurons in your mind. The brains neurons chase down the connections that lead to the properties of a circle, and the circle itself. That sounds like Sutra and Storage. AI is the Hold Grail of computer science, and has proven to be a graveyard for many programmers. Note to self, get real, keep job.

I strongly feel that RDF should be the foundation of Sutra, but it is not an easy fit. RDF's triplets of subject, predicate, and object do not fit well in a DB since the object may be a triplet or a container (bag, sequence, or alternative). I've been reading about RDF Schema, ontologies and OWL to learn how to restrict RDF's mutability to keep it compatible with a relational database. A well defined universe of creators (people), artifacts (files), locations, and events may keep RDF manageable so it can safely be mapped to and XML schema that itself can be mapped to a DB.

I think there are really three key tables, object, attribute, and link. Object but include some concept of type, creators, artifacts, locations and events being the foundation types. The attribute table only needs to handle strings; dates and numbers will be comparable if they are encoded right. The link table will be used to construct a graph of the objects.

I started revising a.g.o's FAQ. It old for starters, referring to GNOME 1.4 designs and bits that aren't relevant. Alex has simplified the language of the site, but it doesn't match the language in GNOME settings, or the code, so the FAQ can be confusing.

Since I've been playing with different GUI's for complex search in Nautilus, I should reuse the ideas in a demo app. I think gsearch can be used to test the code, and provide a demonstration of how Medusa can be used.

I'm recovering from a week long illness. My parents visited for part of the time, and I had to tackle an identity theft incident at work. I didn't do any hacking. I reviewed a few languages; I should do more hacking with python, I can't say I look forward to C++ again.

I built GNOME 2.3 since I wasn't really using my computer. I learned that if you bootstrap with jhbuild, bootstrap everything. I broke my rule of testing (the cough medicine made me do it) as another user–I wont do that again. GNOME 2.3 + Evolution didn't work. I took me a hour to get GNOME 2.2 Epiphany working again. I'm going back to having a test user without fear of messing up my own work environment.

I fear that Time Inc will want to do a cover up of the identity theft's discovered at Time Life Inc. Our investigation indicates a bad party has placed orders for more than 1000 customers. As 99% of the customers have AOL addresses we contacted AOL, who will report to the customers and credit card companies that someone is impersonating the customers and placed orders at TimeLife.com. Time Inc. started stalling the moment they learned this incident could go public. I added rules to TimeLife.com to prevent the party from placing more orders are other customers.