Hi, i am in the middle of my thesis and at the moment looking for about 10 people (i wouldn't mind more) to test my program and fill out a questionnaire about it. The aim of the tool is to support people in their search for scientific publications by providing part of the citation graph for each paper as well as an indicator of how good/bad the paper is. A thorough description and the tool itself (written in Java, provided as a jar file) can be found here. The questionnaire is on the same page. Important: the publication database the tool is connected to mainly consists of papers from the field of physics. a sizable number of computer science and mathematics papers are also available. thanks in advance. ps: any comments on the tool would be appreciated, positive or negative. ps2: sorry for the crossposting, i was adviced to put it in here

Just a few comments, mainly about the graphs the tool generates. I couldn't really comment on its usefulness as I'm not a physicist :P

Anyway, it seems like you have some kind of random placement algorithm going on to determine where each box goes. Now sometimes the graphs I was getting caused scrollbars to appear, and although this will obviously be necessary in some cases why is the window size fixed? If the window was resizeable you could put a lot of extra space towards the graph.

Secondly, when you click on a paper in the graph the info window comes up, but then you have to trawl through the sidebar in order to get the subgraph for that particular paper. The sidebar doesn't seem to be in any particular order, so this is a bit of a chore if you've put in a subject with many papers and you aren't quite sure which way all the arrows go when looking at the whole thing. You could maybe simply add a button to the nodes info panel which jumps to its own subgraph, or even just make it so that right clicking on a node jumps to its subgraph. I'm sure that this would take very little time to implement (especially the right-click option), but it would be a big advantage in terms of useability

Apart from that though it seems pretty good, worked straight away with no connection problems etc

hey, thanks very much for the input. that's a good idea with the button on the info panel. i think when i implemented it, i was focused too much on how i saw it instead of stepping back and taking an objective look at it.

yeah the graph placement algorithm is from an external library (JGraph - www.jgraph.com), sometimes it causes a few problems. the window size is fixed, but when the number of nodes in the subgraph is too large, i give the graph more space.

honey, that is some impressive work. The graph visualization provides a lot of indrect information! I also found it to connect flawlessy to the database and give me no problems at all. It is very informative to see which papers reference which; especially if the graph is particularly complex. I found the fitness algorithm to be spot on; it elevated the fitness of the nodes I would have chosen myself.

If I had to offer some suggestions, my first would be to allow the nodes of the current graph (or subgraph) to be sorted, in list form, by criterion such as fitness, age, name, etc. My second suggestion would be to decrease the time it takes for queries and graph generation, if at all possible.

thanks for your reply. i got 11 people to do the survey, not much but better than nothing :)