2010-10-20

Will KPresenter and Gnumeric Please Come Forward?

This is probably one of the few times that I'm wishing that I had Microsoft Office on my computer. (As it happens, as I go to the library for at least an hour every weekday anyway, I just used Microsoft Office there.) Why?
Well, for my latest chemistry problem set, I need to plot a range of data and add a trend line. Although OpenOffice.org Calc can do this, there aren't as many options. It only gives options for linear, exponential, power, and logarithmic trend lines, none of which are what I want. Although the power regression fits well, what I want is a quadratic regression, and this is something that I just can't do in OpenOffice.org, which is really a shame. I remember when testing some distribution that included Gnumeric (I don't remember which one), I needed to do a similar thing then, so I tried to do it in Gnumeric; if I remember right, Gnumeric did offer the option of a polynomial regression line (with the order of the polynomial specified by the user). Score 1 for Gnumeric, 0 for OpenOffice.org. Also, last year, I needed to make a 3D plot (x, y, f(x, y)), which is possible in Microsoft Office Excel. OpenOffice.org, unfortunately, doesn't have this capability, and at that time (I don't know if the situation has changed much now), it couldn't even render an already-created chart properly. I tried to recreate the same chart with Gnumeric, and, lo and behold, it worked perfectly! Score 2 for Gnumeric, 0 for OpenOffice.org.
Also, last year, I found myself needing to create and view many spreadsheets with lots of data (thousands of rows). Although this wasn't problematic per se in OpenOffice.org, it was certainly a lot slower than in Microsoft Office Excel. Score -1 for OpenOffice.org? Maybe.
So what does KPresenter have to do with all this? Well, it's just that in my experience, KPresenter does a whole lot better in terms of usability and ability to create high-quality presentations than either OpenOffice.org Impress or Microsoft Office Powerpoint. That's because the whole KOffice suite is geared towards desktop publishing as opposed to traditional document creation. Score 1 for KOffice, 0 for both OpenOffice.org and Microsoft Office.
Oracle is being rather wishy-washy about the future of OpenOffice.org, which isn't confidence-inspiring either. I would say the only things OpenOffice.org have going for it are Writer and Math. In the near (or not-so-near, I don't know) future, I may supplant Calc and Impress with Gnumeric and KPresenter. And honestly, AbiWord is a pretty good alternative to Writer as well.

5 comments:

  1. I tried to use gnumeric instead of openoffice, due to the fact that I need to load 40M DBF's each month, and that where incredible slow in openoffice (more than 10 minutes). More, if you take in account that gnumeric loads it in 30 SECONDS. Lately, openoffice calc gets way better, taking 2 mins. The thing is, that in gnumeric, decimals gets lost, and later I realize that was global locale settings that affects the way gnumeric interprets numbers. Thats incredible for me, and more, the way to fix it via some obscure global (reeeeealy hard to discover) command-line setting before execute gnumeric is a shame.

    ReplyDelete
  2. @murray: It's surprising that Gnumeric has trouble with decimals due to locale settings, considering that it's supposed to be essentially a clone of Microsoft Excel 2003 (to the point that it even replicated Microsoft Excel's statistical errors, then fixed them even when Microsoft did not). Then again, there's nothing that says that Gnumeric has all these tools via a GUI. Thanks for the comment!

    ReplyDelete
  3. Gnumeric is a nice spreadsheet but spreadsheets are not adequate for real science. When your data goes beyond a few kilobytes, it's time for shell scripts and the specialized tools available in good gnu/linux distributions like Debian. Two good tools for graphing are gnuplot and IBM's data explorer, OpenDX (start here with DX but the practical bits are to this guide.) If you look around your repositories, you will quickly find other specific tools for every discipline. Proper data fitting is a subject which fills books, but here is a nice example. Happy hacking!

    ReplyDelete
  4. @twitter: Please do understand that this is an introductory-level chemistry class. I know very well that in higher levels of science with higher computing demands, of course spreadsheets are going to be inadequate (and replaced by shell scripts). However, for my needs, spreadsheets are perfectly fine (aided by the fact that I don't really know scripting). Thanks for the comment!

    ReplyDelete
  5. Heh, I understand. Use the links above as you see fit and good luck with it all. If you master the nice tools now, they will be your friend when you run into problems like reader Murray has.

    ReplyDelete