Excel versions brief description

Here is my description of the various versions, add your own views and/or disagree as appropriate.

Excel 97

Big upgrade from 95, 16k to 64k rows. dual file format (remember them?), VBA now in VBAIDE not just ‘sheets’. COM add-ins via xla/xll wrapper only. A few things from more recent VBA don’t work in 97 – like non-model forms. Plenty of people have stopped supporting this version now. Codematic still does, not that anyone asks (not in the last 12m anyway). Many of us are still trading off the knowledge we gained working with this version (Biggus?) (mainly because most changes since have been non-fundamental, more refinements than massive changes). And because actually not that many new developers are rushing in, see Dick here.

Excel 2000

VBA upgraded from VB4/5 (?) to VB6 engine – big improvement. COM add-ins for commands, w/s functions still via xla/xll wrapper. Bit unstable. Very common upgrade (poss from 95?) still standard in many large orgs.

Excel 2002 (xp)

Like 2k but more stable, improved error checking features. Automation add-ins now possible via COM direct from worksheet functions (performance uninspiring). Not enough improvements for many orgs to upgrade from 2k, but a few skipped from 95/97.

Excel 2003

Like 2k again but more stable (rock solid in my experience, you can crash it, but its generally pretty tough), easy on the eye graduated UI. Many large orgs have upgraded to this in the last 2 years. A few tidy-up changes from 2002, but probably not enough to justify a migration for many orgs. XML beginning to be worth using, allowing server access without instantiating Excel (server Excel not recommended by MS). Enterprise management features now working pretty well (policy stuff). This is my preferred version, a good vintage you might say.

Excel 2007

I havent used this for fee paying work so can’t really comment in depth. Initial reactions were disappointment, the crazy UI of course was the biggest, although there seemed to be lots of great new features I couldn’t find them within my frustration threshold. Maybe I classify as too busy/impatient/easily frustrated to migrate. The other big disappointment was the lack of .net integration. I really thought the framework would have been distributed as part of the O2007 install, which would have really opened up the market for 3rd party .net add-ins. And might have helped me recoup my investment in C# learning. (bitter? me?). I kinda expected performant .net UDFs (I havent tested if they are any better than 2003, they might be) as well as realistic .net component deployment. XLM is still going strong though, well its still there, just. I hope we get a usable .net story before XLM is retired.

Excel 2007 uptake forecast

I can’t see many large orgs that recently went to 2003 moving to 2007 in the next 2 years (thats a 3-4 year lag). Those left on 97 are already unsupported so unlikely to migrate to 2007. Many large orgs only went to 2000 2-4 years ago (ie a 3-5 year lag) they wont be rushing to 2007. Most personal users will get 2007 by default with new pcs. Smaller orgs may not be in a position to benefit from the powerful integration features in O2007, or may not have the server infrastructure. So it looks to me like medium sized orgs, and maybe those large orgs with tired 2k installs may make the move, plus small org and personal new pc buyers, will likely be the market. What do you reckon?? Anyone got firm data either way? (If not just make it up like I have!).

The improved integration with Sharepoint and other server stuff may well be the source of pressure to migrate to 2007. This server side influence is a bit of a new thing so I’m really (really really) interested to see the impact. This, I think will tell us a lot about our markets future.

I’m looking forward to see how things pan out over the next 12m, as far as I am aware MS are pleased with 2007 uptake so far.

What were the key features (good and bad) for you in each version? (Don’t bitch about 2007 too much I have another post about that coming up soon ;-))



8 Responses to “Excel versions brief description”

  1. sam Says:

    10 features that didnt improve in 10 years (97 – 2007)

    1.Data Validation
    2.Table (Data->Table….)
    3.Solver and GoalSeek
    4.Text To Columns
    5.Names (very few changes)
    6.Custom Views
    7.Paste Special (only one addition – Column width from 97 to 2000)
    8.GoTo Special…
    9.Advance Filter
    10 Charts (No new charts… Unbelievable !)


  2. Simon Says:

    does this mean you think the UI has improved??

  3. Harlan Grove Says:

    re Sam: MSFT isn’t fully responsible for Solver. Frontline Systems isn’t going to left MSFT give away free (as far as Frontline Systems is concerned), full-featured versions of its main product. And having seen just how long it took MSFT to get linear regression right (and now SCREW UP chart trendlines, which used to be one of the solidest bits of Excel), and seeing how they still can’t get MOD or GAMMALN right (i.e., limits on valid results or number of digits of accuracy), I’d really rather prefer they didn’t attempt to write a Solver clone themselves.

    As for Paste Special, I could have sworn XL97 didn’t have nearly as many options as later versions. IMO, OpenOffice Calc’s Paste Special dialog, which uses check boxes for format/formula/value/etc and RETAINS the previous settings during same session, is arguably much more efficient.

    GoTo Special – agreed! It’d be useful to have a Custom… option in which we could enter a formula that could be evaluated, like with conditional formatting.

    Advanced Filters – it seems to be tied to Excel’s D… functions (DCOUNT, DSUM, DGET, etc.). Way, way back in Excel 2.x, they were just as good as in Lotus 123 Release 2. But then Lotus produced 123 Release 3 in 1989 with huge improvements to its @D… functions, but MSFT let Excel’s D… functions stagnate, and Advanced Filters with them. At times I wonder whether MSFT couldn’t because 123’s new approach was part of its DataLens pattents, or maybe MSFT just didn’t care. But then for a few versions we had SQL.REQUEST, and Excel would have been clearly superior if it hadn’t turned out SQL.REQUEST had memory leaks. Now there’s no more SQL.REQUEST bundled with Excel. Seems to me either MSFT is indifferent to providing anything more than archaic (read: nearly 20-year-old) ‘database’ functionality in worksheets, or it expects any database functionality to be implemented in code.

    For me, the big disapointment is the stagnancy of worksheet formula syntax.
    – It’s still not possible to load multiple files with the same base filename from different drives/directories at the same time in the same instance. I can’t think of any other Windows application that can open multiple files that shares this limitation.
    – External references are still the moral equivalents of DDE links, with static pieces that require very crude hacks to vary path, filename, worksheet or range components dynamically. [And operationally, external references in Excel remain MUCH SLOWER than external references in either 123 or Quattro Pro.]
    – 3D references are still syntactic sugar, available only in a handful of functions. It SHOULD be possible to use INDEX and OFFSET with 3D references [it SHOULD also be possible to use 3D references AND multiarea ranges in COUNTIF and SUMIF]. And there’s no way to pass 3D references to VBA as arguments to udfs.
    – No array versions of AND or OR functions, so record-wise conditional operations require (array_comparison1)*(array_comparison2) for AND and ((array_comparison1)+(array_comparison2)>0) for OR.
    – No quasi error values #BLANK! or #MISSING!, the former which could be returned by cell formulas but treated the same as if the cell were blank, the latter serving the same role as missing values in stats packages.
    – Not making the informational XLM functions regular worksheet functions. [Warning – sarcasm: no, MSFT just had to make DEC2HEX, HEX2DEC and the Bessel functions built-in functions, but who needs GET.CELL or GET.WORKBOOK functions in worksheet formulas?]
    – CELL function’s inability to return anything for CELL(“Filename”,..) when the file hasn’t been saved, even though CELL(“Address”,ref) *CAN* return Book# and worksheet name along with range address when ref is a reference to a cell in a different worksheet.
    – No simple means of returning workbook or worksheet names. The whole reference syntax remains mired in how they did it in Excel 4 workspaces.
    – No worksheet text functions that operate right to left, making it much more difficult than it should be, e.g., to find the last space in a string.
    – No way to turn off @#$%&*! automatic date interpretation other than giving cells the number format Text.

    Yes, a lot of this could be addressed in VBA. At which point Excel is a grossly excessive grid control.

  4. Dennis Wallentin Says:

    Simon et al,

    As for Excel 2007 I have to say that I like the Ribbon UI. Not because it does not support developing (which is a shame) but for making tools more available to the end users.

    I share the disappointment when it comes to functions. A hobby project for me is GNumeric and MSFT should take a look on some of the functions that are shipped with GNumeric.

    However I have the impression that the calculation engine have been improved with every new version which is good.

    Harlan – Does all the XLM4 functions still work in Excel 2007?

    I’ve forgotten when MSFT officially removed the support for it but it seems that it’s still available.

    The QueryTable works better in .NET then in Excel because the object is now part of the managed object model for Excel 2003 and 2007. We can therefore use ADO.NET (among other features) and create connections through the .NET Wizard. The ListObject and the Chart object are two other objects that can be used in a similar way as with the QueryTable object. The DDE has been replaced with RTD but only through managed/unmanaged COM add-ins for 2k and later versions.

    There are some improvements made to connect to SQL Servers (2k and 2k5) which I like.

    In general I believe that the lack of improvements are due to
    a) all larger customers does not ask for them and
    b) Excel has reached an impressive age where the core parts of it cannot easily be revised.

    Kind regards,

  5. Jon Peltier Says:

    I think the lack of improvements is due to:

    a) Excel is no longer considered the killer app, it’s very mature, and newer technologies (sharepoint, onenote, etc.) are sexier.
    b) Much of the effort on Excel 2007 was spent incorporating new Office stuff (Ribbon, new shapes), so there were no resources left for the Excel-specific things.

  6. Marcus Says:

    Merged cells… aaargh. Sorry. Just had to get that out.

    I’m amazed tha MS haven’t taken Excel to the next level in database connectivity. More specifically as an interface to datawarehouses and OLAP sources. Yes I know it exists but it’s still not compelling enough when compared against other BI client tools.

    On the other hand… are we simply asking Excel to be more than it’s intended to be. I’m sure most of us have witnessed people typing adocument in a spreadsheet. Is this perhaps the same principle? Call a hammer a hammer?

    Regards – Marcus

  7. Simon Says:

    Harlan, great list, thanks – you mean you don’t use DEC2HEX everyday? (me neither)
    Dennis XLM still works fine in 2007 AFAIK, they don’t make it easy to find macro sheets but I’m sure they work. There are rumours it is approaching End of Line. I had some XLM work less than 12m ago. All that ExecuteXL4Macro stuff still works in 2007. good points on lack of improvement.
    Jon good points too, especially sexiness, looking at OfficeZealot, you could be forgiven for thinking that MS Office is Sharepoint and Sharepoint is Office.
    They did do multi-threaded calc in Excel, I bet that wasn’t a 10 minute job! And the big grid (edit replace long with LONG64). But I know what you mean about Excel-specific.

  8. Harlan Grove Says:

    Me, I suspect MSFT marketing had figured out that few would have upgraded to beribboned Excel without providing the big grid.

    BTW, what problems have any of you seen importing or linking Excel 2007 ranges into Access 2007 as tables now that Excel provides a quantum level more columns than fields/table Access could handle? Not that I can think of a good reason (perhaps aside from crosstabs) for database tables to have thousands of fields.

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: