I recently wrote a macro so that Word could call an AI API to do AI-assisted translation, works like a charm.
The main advantage of office 2003 of course is that it's the last office without activation and other crap: you pass the serial and own it for life, it won't bother you again.
I wantwd to only use 2003 but after the 10th time I argued with a person that sent me a docx for editing I gave up.
The issue is that roundtripping between Office 2007+ and Office 2003 is unreliable and will often result in corrupted files.
Using Office 2003 (with Compatibility Pack add-on to open xlsx and docx) is ok for isolated work but can be unreliable for collaborative back & forth editing depending on what features are used. E.g. cell colors used in Excel 2007 xlsx get corrupted in Excel 2003 xls.
https://learn.microsoft.com/en-us/openspecs/windows_protocol...
That's actually not true, Office has had activation since XP (2002), so 2003 is included in that.
It works fine if the user is ok with the features from 2003. E.g. Excel 2003 is limited to smaller spreadsheets of 65536 rows by 256 columns but Excel 2007+ can handle larger worksheets of 1048576 rows by 16384 cols.
I also recently used Excel's new LAMBDA() function which was introduced 2020. In earlier versions, it required writing a VBA UDF to accomplish the same task of assigning a temp variable with a ephemeral value to calculate on intermediate values. VBA is a workaround but LAMBDA() is nicer to use because Excel will throw up scary security warnings whenever the xls file containing VBA macros is opened.
I might be able to get by with Word 2003 more than Excel 2003.
A lot of normal users of Excel are not users of database software like SQLite or MS Access. That's too cumbersome. E.g. they download a csv file that has ~100000 rows (which really isn't that "big") and clicking on it gets them an instant visual grid as a GUI in Excel. Slicing & dicing and pivoting data is way faster in Excel than coding SQL WHERE GROUP BY statements. I've commented previously on why databases are not substitutes for the workflow ergonomics made possible by spreadsheet tools : https://news.ycombinator.com/item?id=30987638
It's similar to reasons why Python/R or Jupyter notebooks are also not substitutes for Excel for the typical users of Excel.
The low row count of 65536 in Excel 2003 was just a legacy limitation of 1980s 16-bit computing that was carried over into 32-bit computing for many years for backwards compatibility reasons. Spreadsheet users don't really want to switch to databases or Python just to get more usable rows than 65536.
> The scenario of "I just sent you an xlsx where the rows highlighted in red are problems and if you can just add your notes to column K, that would be great. Thanks!" -- is not easy in other tools that are not spreadsheets.
There are no words to tell how much I hate that!!! ;-) Users are too creative. Some will merge some cells and not others and boom the file can't be properly sorted anymore. Many will use font color, font weight or background color to mean wildly different things, which is very difficult or impossible to sort or do a sumif over. Others will add footnotes, because why not?, and links to spreadsheets that never leave their own device, or mini-blank-rows for spacing and general layout, etc. It's completely insane.
Your side lost completely. Pop a signal flare or build a bonfire, maybe someone can rescue you from the island you've been living on since the war ended.
XLookup sure is useful but again you can probably replace it with a combination of vlookup and hlookup (or index match).
Regarding the size... If you're dealing with huge spreadsheets it's really better to use a proper db. Or even manipulate data with sqlite. sqlite can query xlsx files directly (with an extension), it's extremely fast and stable.
And whilst you can work around lack of XLOOKUP or SUMIFS using the older functions, again it constrains how you lay things out (eg VLOOKUP needs you to presort your table by the lookup column if you don’t want an exact match) and this can often result in sheets which are much more unwieldy and slow to calculate.
I remember Outlook clipping the last character off the email subjects, for example. Might have been Office 2010.