27. April 2008 03:24
Some time ago I noticed that Fredrik e. Nielsen had posted a link in a Norwegian debate to a website comparing spreadsheet formula interop using ODF. The article is from 2005 comparing formula interop between OOo Calc 1.9.117 and KSpread from KOffice 1.4.1. The article is interesting since it highlights one of the more serious problems with lacking spreadsheet formula definitions in ODF. Some of the pictures in the article are missing and because the article is 2.5 years old, I thought it'd be interesting to take it for a spin again and see what has happened since 2005 in terms of interop between the two major ODF-implementations. I have done exactly the same as in the original article and have additionally added a bit of research to see where the problem really lies.
What did I do?
Well, on my brand new ubuntu 8.0.4 installation I installed KSpread 1.6.3 in addition to OOo 2.4 that came pre-installed with the system.
I created a spreadsheet using OOo 2.4 Calc with the data from the original article (formula OOo.ods (7.58 kb)
I then tried to open it using KSpread. This is what it looked like:
As in the original article I modified the formula to fit in KSpread and the result was:
The file s available here: formula KSpread.ods (5.44 kb)
When saving this file and opening it in OOo again, this was the result:
So there has actually (pheew) been some improvement in spreadsheet formula interop for applications using ODF Spreadsheets since 2005 ... thank God! At least now OOo is able to show the formula created by KSpread.
To take a more deep look into what was the cause of the problems, I added some information to the original spreadshee. The result is here: Since OOo can read the formulas from KSpread, I have opened the file using KSpread to demonstrate the problem:
The file is available here:(formula OOo exp.ods (9.30 kb)
So what should we conclude from this very basic test? Well, you tell me ... but at least, when someone next time tells you, that lacking formula spec in ODF is not a practical problem but only a theoretical problem ... please tell them that they are wrong.