darrenyeats;619207 Wrote:
> If over the network you open a spreadsheet from a Windows file server
> and it's different to the same spreadsheet opened from a Linux file
> server, that's a bug.
Not necessarily. I have extensively tested spreadsheets and can find
subtle but real differences.
Of course, you have to have a system capable of resolving those
differences in the first place - you're simply not going to see the
difference in two Excel spreadsheets if you're comparing them on, say,
an Asus Eee and a Dell netbook. Neither machine is capable of bringing
out the best in a spreadsheet, so it stands to reason that neither
machine will show differences.
At least, my testing has failed to show any.
Second, you have to compare 'apples' to 'apples.' It proves nothing if
you're comparing the payroll records of a small start up to the sick
time of a large corporation. They're too different to make any
meaningful judgement.
In order to adequately test spreadsheets, I've found it's important to
precisely match formulae and records length. But even after you have
'level matched' you have a number of further steps - the most obvious
being in the areas of power supply and cabling. I disconnect the stock
power supplies on the units I test, and use a purpose built linear
power supply with balanced XLR outputs.
Cabling adjustments include an oxygenated mouse cable - I don't claim
to understand why this makes a difference, but it is clear across a
variety of spreadsheets and texts. Since I apply it equally to all
tested documents, I figure it can't hurt.
With all the above done - and remember, the point here is not 'tweak'
endlessly; it's simply to get the best spreadsheet results I can - I
have found network issues can make a huge difference.
For instance, I have stored a favorite spreadsheet of mine ('small
tools & misc. supplies, March '010') on both a Win 2003 box with a
stock power cable and a much larger Linux server (RHOS '09 with the
Friedlander extensions). To me, the differences couldn't be larger.
Off the Win server, 'small tools' is very good. The overall inventory
levels and controls are exactly where they should be - there is the
transparency to the records you'd expect, and a 'rounded quality' that
allows one to get a strong sense of the story the spreadsheet tells. In
particular, I delight in the way 'drill bits, metric' moves within a
defined but expansive range between the 13th and 29th.
The Linux server with Shinyota power cable is something else again; all
the above is true, but a sense of *detail* joins the fray. It's as if I
could drill down through endless layers of detail and there would
always be more to find. Yet I don't feel as if any information is being
kept from me; if anything, it's as if a veil has been lifted.
Could I prove this with an ABX test? Doubtful, but as others have noted
I think that says more about such tests than about what I perceive.
Scott A.
--
satkinsn
------------------------------------------------------------------------
satkinsn's Profile: http://forums.slimdevices.com/member.php?userid=11398
View this thread: http://forums.slimdevices.com/showthread.php?t=86419
_______________________________________________
audiophiles mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/audiophiles
