Microsoft Office 2004: A Product Vision Study
Instant Results Computing
Microsoft (and indeed the whole PC industry) desperately need to drive upgrades of all sorts. Software upgrades are foremost on the minds at Microsoft, but getting people to upgrade their hardware is of great benefit too and indirectly sells software (look at all that stuff that comes bundled on a new machine!). One of the issues that people have with upgrading hardware is that many are starting to feel computers are fast enough. I read that as software hasn't been able to harness the greater power available to good benefit. With the notable exception of games, this is largely true. You've read about Borg Cube, my home machine. It's a real hot rod. My work machine is a 650 MHz laptop that is dramatically slower. Yet, when I use MS Office, I hardly notice the difference. It isn't just raw speed either, available RAM and hard disk space on newer machines is much larger and there are bound to be opportunities to take advantage of this. I've coined the term "Instant Results Computing" to refer to features that utilize the extra power of newer machines.
Instant Results Computing means taking available extra computing power and using it to enhance productivity or provide a nifty feature. I prefer to place automatic features in this category. If you had to manually initiate an action to use the power, it isn't "found money". MS Word's ability to spelling and grammar check in the background while you work are excellent examples of Instant Results Computing features. Let's look for some more.
We talked about the Concept Encyclopedia and the desire for Google-like search engine functionality on the local machine. This costs a fair amount of cycles. With Instant Indexing, we would be continuously keeping the Encyclopedia and Search Engine up to date with what they need to do to find the current document. By the way, the old manually updated table of contents and indexes should have "background recalculation" in MS Word and should never need manual updating.
Much as Word underlines potential misspelled words and grammar problems, Instant Links would add hyperlinks to documents on a proposed basis in real time as you type. A blue underline would be used, as this is a standard hyperlinking convention. All of the terms that appear in the Concept Encyclopedia would be listed, but in addition, the system would be on the lookout for candidates to add to the Encyclopedia. These would be marked with a blue wavy line. In looking for new hyperlinking candidates, we want to find terms and concepts that are relatively unique. Proper names, often identifiable because they come up as misspelled words which the user says to ignore, are good candidates. Words that have a low frequency of use are probably interesting, as are words that have higher frequency in the document than in normal English usage. For example, this web article mentions MS Office and User Interface much more often than these concepts are found in "average" documents.
Instant Formula Correction
We don't really need the extra hardware for these features, but I like the buzzword so I'll keep it here. Word does lots of whizzy things to automatically identify and correct errors, while Excel does very little along these lines. I understand Excel XP is a little better along these lines, but I'll bet a lot more is possible. Let me give some examples of some automatic changes Excel could make that would make spreadsheets more robust.
Our first example concerns the use of the SUM-style functions (AVERAGE is another such). SUM is probably the most common function used in spreadsheets. A common error is to create a sum of a column of numbers and then to add a row at the bottom of the column. If the sum isn't written to include a blank line above and below the column, it's easy to develop an error where the new row isn't part of the sum. This isn't obvious unless you check the model very carefully. Why not have Excel expand the sum range to take in one row above and one below as long as they contain non-numeric and non-formula contents, which the sum will ignore?
Another example involves entering formulas. It has been a thorn in the side of Excel users for a long time that you have to remember to type "=" or the system will not consider what you enter to be a formula. Just typing in "2+2" gives the result "2+2" instead of "4" because of this. Why not have Excel look at what was typed and decide if it could, in fact, make more sense as a formula?
A final example that happens to me all the time. Enter a column of numbers. Place a SUM at the bottom to yield column total. Now, in the column immediately to the right, we want to compute what percentage of the total the number on the left is. So, we point to the cell on the left, enter a divide slash, then point to the column total. We then format that as a percentage. So far so good. But if we failed to make the column total an absolute reference, and then we copy down, we have a bunch of divide by zero errors. Excel should spot this and offer to fix it.
There are many more heuristics possible for this sort of thing. I've just given a few examples. Done properly, this could be an incredible labor saver and a real nice feature for Excel.
All material © 2001-2006, Robert W. Warfield.