On Sat, 2005-27-08 at 10:46 -0600, Lakin Wecker wrote:
> Moving this email to the [email protected] mailing list:
> 
> I work for a small company.  I've recently convinced the boss to switch
> over to OpenOffice.org and then subsequently to a Free Software desktop
> as their main office suite. They use a number of different Excel
> spreadsheets to do most of their work throughout the day.  As such, Calc
> is the critical application, and the subject of this email.
> 
> After some careful consideration I decided to try out the beta OO.o
> 1.9.x, as it will hopefully be the released version when they fully
> switch at the beginning of 2006.
> 
> I installed the beta version onto two of their computers(Windows 2000)
> and tried two of the spreadsheets.  I was very surprised at how slow
> Calc was at opening each of the spreadsheets(See timing info). After
> some investigation I noticed that the equivalent Calc format
> spreadsheets were significantly smaller than the equivalent Excel
> Spreadsheets; this is great for people looking to save disk space.
> However these employees open and load many different spreadsheets
> throughout the day, and this wait time will be quite annoying for them.
> 
> 
> Here are the stats:
> All of the load times are approximations as I used my stopwatch to time
> them.   On the windows machine the OO.o quick start was loaded. I timed
> from when I double clicked on the file to when the program presented the
> screen with all of the numbers visible and the UI was usable (scrollbar
> worked).
> 
> Computer specs: AMD Duron 1Ghz 512 MB Ram, Windows 2000.
> 
> File #1: 
>         Excel Format Size: 7.05MB
>         OO.o Format Size: 1.68MB (1.77MB with XML optimization off)
>         Excel Load Time: 4s
>         OO.o Load Time: 1m40s (either format).
> File #2:
>         Excel Format Size: 192Kb
>         OO.o Format Size: 64Kb
>         Excel Load Time: < 0.3s
>         OO.o Load Time: 10s. (either format)
> 
> I tried the files on my Ubuntu Laptop which is a 2.0GHz AMD64 machine
> and got the following load times with OO.o 1.1 and Gnumeric 1.4.2:
> File #1:
>         Gnumeric Load Time (Excel format): 5s.
>         OO.o 1.1 Load Time (Excel format): 26s.
> File #2:
>         Gnumeric Load Time (Excel format): < 1s.
>         OO.o 1.1 Load Time (Excel format): < 2s.
> 
> 
> I assume these discrepancies are the result of caching formula results.
> The assumption is that Excel stores the previously calculated results of
> the formulae in the spreadsheet and uses these values at load time,
> while OO.o ignores these and fully recalculates the formulae on each
> load?  I haven't been able to find concrete proof of this, can anyone
> confirm this is the case? I tried looking for an option to force OO.o to
> use more disk-space and save on load times. I found: "Size Optimization
> for XML Format" and un-checked that option, but did not notice much of a
> difference(9k in file size, nothing for load times).    

According to someone on the [email protected] mailing list, OO.o does
indeed cache the formula results.   If this is the case, then something
else is going on.


> I'm curious to know if there are plans to address this for the 2.0
> release? If not, I am willing to put in some work on this, but would
> need some direction as to where the relevant processing takes place.

To clarify this point: I'm more interested in whether this is a known
issue.  If not, I don't mind doing some profiling and digging to find
out what is the problem.  I just don't want to waste any time on
something which is already being worked on, and/or finished.

> The concept is that there would be a check box which switches between
> disk-space optimization and load-time optimization.  At the very least,
> when using Excel spreadsheets, the OO.o loader should be similarly fast
> at opening the files. 
> 
> Sincerely,
> 
> Lakin Wecker
> 
> 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to