Linux Desktop for university staff

[David Curry]

Jason Powers wrote:

> I don't mean to be contentious, here, but
>
>> I think "perfect alternative" to excel is a bit of overstatement 
>> regarding oo.o calc.  Calc is limited to 32,000 rows whereas excel 
>> can handle larger files and calc does not have as many built in 
>> functions for users to draw on (date manipulations such as 
>> determining the time between two dates come to mind). 
>
>
> Do your users runs a lot of 32K+ row excel spreadsheets full of 
> formulas? You might want to look into training them for postgres or 
> something. We had one user running a spreadsheet up around 15K rows 
> and I kept screaming at the DBs to load it into ingres (our dbm at the 
> time) - having to restore it from tape twice a week was not fun, and 
> we kept getting her new equipment, more memory, etc. All to run this 
> one spreadsheet that would take her machine down periodically. Maybe 
> our newer 3ghz PCs could handle it, but I'm still happier having it in 
> a database.
>
>> The "R" statistical package (http://www.r-project.org/) is a good 
>> substitute for S" and should not be overlooked in your list of software.
>>
>
> We make this! Rather, I work for a guy who designed it (Dr. 
> Gentleman). The stats here who start using it invariably stop using 
> SPlus, saving us thousands of dollars in annual licensing fees. It 
> runs most splus code without much tweaking, so it's pretty easy to 
> migrate. Get this - they like it better. We don't have to twist their 
> arms or anything. One or two of them have gone so far as to stop using 
> SAS in favor of it as well, though I expect this is a matter of them 
> not using SAS for big stuff anyway.
>
>> Gretl (GNU Regression, Econometrics & Time Series Library) is another 
>> application worth mentioning to the university crowd and can be found 
>> at http://gretl.sourceforge.net/
>>
>
> Now I have to try this one out, thanks.
>
> Jason Powers
>
Thanks for the feedback on R, Jason.

I don't know that I have any disagreement with you regarding large data 
sets subject to repetitive use by multiple users who have reason to 
believe they will realize continuing benefits from learning how to 
maintain, and generate reports from, the data base.

But, there are one heck of a lot of large data intensive situations 
where spreadsheets are quicker, easier, and less costly to the user than 
a database implementation.  SOHO operations may encounter them more 
frequently than large institutions, but I suspect that quite a few 
university professors pursue research studies that generate them as well.