Here is a partial and dynamic "wish list" of Linux programs proposed by members of the SEUL-sci mailing list. Comments, suggestions and pointers to programs filling these niches are always welcome, as are new items for this list.

Data Acquisition
     Given its ability to run on a wide variety of hardware, Linux should be an ideal platform for data acquisition, especially in a networked enviornment. What DAC packages are there presently available for Linux?

'Why Linux for Science'
     What makes Linux a good operating system for scientific uses? Why use Linux over Windows9x or NT?

Meta-Analysis Tool (aka 'data-stealer')
     This would be used to facilitate reuse of data from several datasets taken from the literature. This program would take a scanned image of a scatter plot, and allow the user to obtain an approximation of the data in the plot. Some code to do this in visual basic already exists and it should be fairly adaptable.

R-Gnumeric linkage
     A linking program between the R Statistical package and the GNUmeric spreadsheet (for that matter, any Linux spreadsheet) would be useful.

GUI-Based QBE Grid
     This program would get information from a database, expose the tables and fields within that db, and allow the user to create an SQL query while permitting but not requiring hand-coding the SQL. Loath as we are to admit it, the Access database has a pretty good QBE grid ... which is good because MS Query's QBE left much to be desired. Related projects worth examining include gASQL and GNOME Transcript

Generalized Analysis Manager
     This is something that's been under consideration for some time now, more out of need than anything else. However, no serious development has yet been started. This is not so much a program as it is an interface between existing software packages: a graphing utility (say gnuplot), statistical analysis package (R?), database for data storage, and spreadsheet for fine data manipulation. The app itself would log all data transactions, graphs and analyses. This latter function would be useful because it would allow the researcher to modify their assumptions or approach somewhat (ie. remove outliers) and rapidly re-do the same analyses and graphing done previously.

Please contact Pete St. Onge (pete@seul.org) for additions