Hello everybody,
this is good news. I will add some important strategic suggestions in a
later post.
This is only a short comment.
TOC
===
1.) AUTOMATIC DATE RECOGNITION
2.) TEXT-vs-NUMERIC
1.) AUTOMATIC DATE RECOGNITION
==============================
Regina Henschel wrote:
Hi,
Niklas Nebel schrieb:
Hi,
[...] quarterly reviews for each project. There is now a wiki page at
http://wiki.services.openoffice.org/wiki/2008_Q2_Review_of_Spreadsheet_Project
to collect issues, and on May 5th there will be the first IRC meeting
(check if all important issues are raised and sort out the
unimportant ones). If you think an issue should be included, please
add it to the wiki page.
I've added issue 87999.
kind regards
Regina
Indeed, issues 87999 (11 votes and a lot of duplicates) and 33723 (31
votes) are related and cause a lot of pain. Both involve more or less
the automatic date-recognition mechanism in OOo.*
* http://www.openoffice.org/issues/show_bug.cgi?id=87999
http://www.openoffice.org/issues/show_bug.cgi?id=33723
I strongly support modifying this behaviour in a short timeframe and
implementing a date input-mask, where every user can accurately direct
OOo as to what input is a date (e.g. only dd.mm.yy, and not xx.xx)
2.) TEXT-vs-NUMERIC
===================
Issue 5658 (http://www.openoffice.org/issues/show_bug.cgi?id=5658)
covers this scenario.
My comment could easily slip into bashing, so my apologies IF this happens.
PROBLEM
How should as-text-stored numbers be handled?
Well, from all the possible handling-options, my personal impression is
that OOo has chosen the worst one: It *silently* ignores these numbers.
The following problems exist:
- NO user notification that something went wrong
[aka Raise Error / Flag result]
- NO easy user action to convert text to numbers
- NO easy way for user to detect which numbers
are stored as text
So, the proper implementation should:
1.) convert text-to-numbers automatically
2.) raise error / flag results where such text occurs
3.) offer easy recognition and conversion of text to numbers
ALL 3 mechanisms should be implemented.
ADDITIONAL COMMENTS
======================
The text-to-number issue needs some additional comments. Please note,
that I am the guy with the strong-typing in spreadsheets request
(http://www.openoffice.org/issues/show_bug.cgi?id=79924 - I'll return to
this request in a later post).
Unfortunately, separating numbers from strings is a very *artificial*
separation. It is NOT the way spreadsheets and humans function. "3" is
always "3", independent of it being addressed in a numerical or textual
context.
This distinction is the way classic programmers think and programming
languages like C/C++/java function, BUT not the way spreadsheets should
function. The 2 worlds are very distinct ones and - unfortunately - the
early programmers did NOT understand much of spreadsheets (and probably
this did not change much over the years).
Therefore, I would completely drop this distinction in spreadsheets:
this is only important for how the data is stored and handled
internally, BUT this is *nothing* the user should care about. It is more
like a formatting option (BUT - strictly speaking - it isn't a a very
good one, because it misses a lot of issues - so users will apply their
own formatting anyway).
On the other hand, it does make perfect sense to distinguish: scalar
numbers (stored as number proper or text) from currency, from date and
from other units (including time units). A strong typing model is
needed, not a formatting option to display a number with a currency
symbol. Instead, formatting a number as text or "number proper" is
actually a *formatting option* - in strong opposition to the true units
I mentioned earlier.
I will discuss typing more broadly in a new post - as it is a very
important and strategic concept.
Sincerely,
Leonard
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
