Numbers: disappointment of the year :(

I was very excited this morning when I have read Numbers have been released. I have been searching for an Excel replacement for quite a while now, and thought my dreams have finally come true.
Excel is not Universal yet, and won't be Universal for almost half a year. And everyone knows Mac version of Excel is crippled (I remember running Windows version of Excel 2003 under Virtual PC because it was faster than native Excel 2004).
Alas... Numbers is not the replacement I was so hoping for.
It is absolutely useless - and I am not sure Apple will ever fix it by the looks of things.

1. There are no error bars in Numbers. Not in the bar chart, not in scatter plot. Numbers can calculate stdev - why not make inserting error bars easy? Even old versions of Excel could do that.

2. Scatter plot implementation is simply a joke - and a really, really, really bad one. Yes, it is THAT bad.
First of all, there is no easy way to plot several sets of "Y" data against a single set of "X" data on a single chart. Instructions in Help specify that two columns of data have to be used for each data series. No, duplicating my "Time" data for every one of my 12 experiment runs is NOT an acceptable option. I have tried to drag additional data columns to the chart, but I could never understand the results of this action.
Second, the data markers are not resizable - and their default size is nothing short of gargantuan.
Third, there is no way to make a spline line through all the data points.
Fourth, Numbers incorrectly names data series. When I select two columns, it is quite obvious that the first column is "X" data - and it is plotted accordingly. Unfortunately, Numbers labels the data series by the name of that first column. Who would ever want to do that?
My fifth and final complaint about scatter are the SLOPE and INTERCEPT functions. Knowing slope and intercept for a set of calibration data is really useful - as long as I also know the corresponding R2 value, and can visually assess the quality of the linear fit on the plot. In Excel it is a two-click operation. In Numbers it is impossible.

3. There is no real inspector for data series. No central place to rename them, change the scope, etc. Even very old verisons of Excel have that.

4. Dragging and selection are seriously mis-designed - and need a fix NOW. Yes, it is THAT annoying.
When hovering over a table, the cursor changes from a "cross" selection cursor to an "arrow" move cursor if it is in the top 25-30% of the topmost table cell. This requires the user (me) to be extremely careful when making a selection. In half the cases I end up repositioning the entire table on the page, instead of selecting the top five rows. After a few times this behavior becomes really, really, really aggravating.
On the other hand, when I want to drag data, it takes some mouse hunting to find the position where the column is actually draggable.

At this point I have stopped my evaluation of Numbers.
I am not talking about fancy curve fitting and data processing a la Origin, Sigmaplot, or ProFit here. I am talking about essential features and usability.

It is really amazing that the same company that has managed to produce Keynote (a really marvellous piece of software) and iLife has now produced such an unadulterated piece of crud. It is simply amazing.

As much as it pains me, after this I am considering a move back to Windows after using OS X for 5 years. Windows can be annoying and prone to entropy, but at least there is a proper Excel version - and also Origin and Sigmaplot. I have yet to find ANY decent Excel alternative for OS X, and that piece of crud Numbers is might have finally broken the proverbial camel's back

iBookG4user said:

Have you tried using NeoOffice J?

Click to expand...

Yes, I know what NeoOffice is.

I would never want to use OOo even when it is running native on Windows or Linux - let alone that kludgy Cocoa/Java port. Ugh.

It runs even slower than Excel 2004, it does not have all of the Excel's features, and it is butt ugly to boot. Oh, and last time I checked, it does not export its charts to clipboard properly, so that when they are pasted to Keynote, they are transformed to bitmaps.

Did I mention it also looks butt ugly?

Thanks for the suggestion, of course.

c-Row said:

Except for #4, I never heard of any of these "essential" features.

Click to expand...

If you are a scientist, having error bars in your data analysis app is absolutely essential.

I cannot prepare a chart with no error bars - because the error bars are as essential for understanding the data as the values themselves.

As far as my complaints about the scatter plots go, again - my use case is quite typical for almost any science user. We want to do scatter plots, and a lot of them. With error bars and spline guidelines for easy readability. We want to be able to do at least linear fits to our data - easily. We want to be able to toggle between linear and log scales for the Y axis. We want multiple data series on every plot, with fully customizable markers, traces, error bars, and axes.

It is understandable that you might not have thought these features to be necessary - you are probably not a scientist.

However, it is INEXCUSABLE that the design team at Apple did not anticipate such a typical use case, and has made a few absolutely idiotic and blockheaded design decisions.

And in this case it is really a shame because when I have first started it, the app looked so pretty - and there are plenty of nice touches in it everywhere. Sadly, it is useless

And speaking of useless, Pages has become a really neat app - but it still cannot do endnotes, and it still saves in proprietary bundle format - which is not parseable by any existing bibliography application. Thus, Pages cannot be used with either Endnote or Bookends... Which makes Pages completely useless for any scientific writing...
Thankfully, there are such great apps as Mellel and Nisus Writer.

netdog said:

It's not meant to be an Excel-killer. It's meant to be Excel for the rest of us. As someone who used to train people in Excel, rest assured for 99.9% of us, Numbers is just fine. For those of us who do financial analysis and the like, and require very sophisticated formulas and macros, Excel is still the only game in town.

Click to expand...

I was not even looking at Numbers for any kind of really sophisticated analysis. I have Origin for that (Windows only and running in Parallels, but what can you do).

Here is what I wanted Numbers for:

1. Do rough and quick examination of raw data (import CSV, do a quick plot, do a quick linear fit to the first 10 data points, print, discuss with colleagues).
2. Do simple cleanup, normalization, or unit conversion operations on data. For example, I have an instrument that produces its timing data in minutes. I will, obviously, want that in seconds. Or I would want to normalize a set of data (it boils down to dividing every cell in one column by every cell in another column).
3. Make nice-looking plots and charts for presentations, posters, and publications. For all its power, Origin makes decidedly ugly plots.

As you can see, it is hardly rocket science - I do have a real data analysis program for actual rocket science.
Yet Numbers cannot do those simple things (which I could EASILY do in Excel 97 for Windows 11 years ago - with at most two clicks of a mouse).
Extremely disappointing.