Those are good additional points Yogia.

I guess reliability comes under weatherabilty. One thing I learned early is that text books and handbooks are not reliable they have many typo's. (Or at least several of the ones I have do) Also standards are usually published with a collection of addendums and corrigdenums.

I have a preference for at least two independent sources. Independent sources are important, not very helpful when text has an error, so look in a handbook and then find same formula, but then also note handbook references same author as textbook.

I prefer books which show derivation of formula, unfortunately many formula are also empirical and tests to clarify not exactly readily available to most users. The next best thing to two references are tried and tested design handbooks. Books which have been in use for 20 to 50 years or more have a good chance of being correct, though the newest edition may not be.

What I am leading to is that simply down loading a workbook and commenting on it is not of much value. Really need to know that many people have downloaded and actually used the workbook for a real design problem. And more importantly that they have confirmed that it is correct. Not simply that the workbook is correct but that the source of the formula was correct.

When pushed for time, I simply use formula from a handbook, and if possible derive afterwards or look for additional references and theory.

When it becomes increasingly easy to simply plug numbers into a program or workbook, for many people the checks and measures disappear. Or even looking in handbooks and relying on data presented such as section properties and design capacities.

Computers don't increase the risk. Its just that if programs have errors in them you get in trouble a lot faster, and a lot more people are enabled to get in trouble.

But it depends exactly on who the audience is and what expectations there are for the workbooks submitted.

At present the workbooks contain simple formula, or simple sequences of formula, no named ranges, no lookup tables, and no drop down lists or data validation. (With some minor exceptions.)

Does the repository keep things simple, simply providing a library of simple functions for users to incorporate into their own more advanced workbooks? Or does it also collect the advanced more complete design solutions? Or split into two sections, one section with the simple formula, another with more complete solutions? Also being international there are matters of differing design standards and units of measurement. Further divisions for filtering and sorting the workbooks.

In 1989 I first used Lotus 123, somewhere between 1985 and then I was introduced to Multiplan. For me it was easier to solve problems in a spreadsheet than mess around writing Fortan 77 programs, especially sequances of simple calculations. But because I was limited by DOS and simple text capabilities I simply adopted programming techniques for setting out spreadsheets. Though I tended to present formula and then the numbers substituted in the formula, then the result. A time consuming activity. By 1996 I gave the idea away, I had windows and improved text capabilities, but the formula were more complex and longer, stretched across the page, too many subscripts and Greek characters (written out). The formula were in codes of practice, so no real need to present. After all what do the input variables mean, look in the code of practice. The calculated variables are not any different. A variable may or may not be dependent. What becomes important is that I need to know the value of some variable to make a decision. The fancy formatting disappears, most design software simply shows inputs and outputs, and excludes the process of calculating those results. My use of Excel and spreadsheets have moved in that direction. More over I am now more likely to write VBA functions rather than create long cell formala, creating a further level of hiding process.

One reason for never adopting MathCAD is that all its demo examples show too much for my liking, and consequently consume a lot of paper. Paper has to be managed and stored. From my view many structural calculations are a waste of time and paper. We know the answer already, it is just ritual and routine, produce calculations to submit to the regulator to get approval. So I aim for maximum calculations in minimum space, and try to keep readable.

What it means is that I and most probably others would have to do a great deal of work to create a workbook formatted with XLC suitable for the repository. Since we have had so long without the facility of XLC, and got by with out ever getting MathCAD.

But maybe a new age is dawning and formatting expressions is back on the books.