Archive for October, 2008

iPhone NDA

Monday, 20th October, 2008

El Reg has a pretty succinct piece on the iphone dev world here.

I’m seriously considering getting a G1 when they come out (unless they get completely slated in the reviews).

I have a couple of phone app ideas, and I reckon Android is the lower risk way to dip a toe in the phone app business. And I can later use my success (or failure) to justify the nice shiny new MacBook I’m after to port the app(s) to the iPhone. Oh and I’ll need an iPhone so I’ll have to give the G1 to one of the kids. Or maybe I’ll keep it if its better.

Oh, and the app is not a spreadsheet, although I see the one for the iPhone seems to have created a minor buzz. And is for sale at the premium price of 8 USD (thats a big premium in a market where most stuff is 1 dollar)

cheers

Simon

New presentation laptop

Friday, 17th October, 2008

I think I have finally decided on a laptop for presenting, and for taking to conferences. And any future Microsoft Office Developer Advisory Council meets if there are any.

Heres the link.

What do you think?

Perhaps we could group together and get a bulk buy discount?

cheers

Simon

Microsoft: We had to discount Office 2007

Thursday, 16th October, 2008

Microsoft admits it had to discount Office 2007 to boost uptake

They have a lot of discounting to go before they reach the OpenOffice price point (and ease of installation – no insulting product activation in OOo).

Of course they could choose to compete on quality, features and benefits rather than price. That would certainly be my inclination against a zero cost competitor. I wouldn’t go for pointless irrelevant (destructive?) product differentiation either though.

Discounting is the classic finance solution to weak sales. Marketing would plump for better eduction of potential customers on the benefits (perhaps by sponsoring the Excel User conferences??). Engineering would go for more/better features. (Its called functional fixation – We did it at college so it must be true)

I’m with engineering here. What about you?

I guess marketing might go for discounting if they felt the product was priced above the customers valuation. But this is meant to be the premium productivity suite. They still have Works don’t they?

Compare and contrast with Apple who have pretty much just increased prices across their whole notebook line up.

(And OOo 3.0 runs natively on Macs as well as the other two main OS’s Windows and Linux)

Just remind me, what is the point of Microsoft Works?

cheers

Simon

Excel headline

Thursday, 16th October, 2008

Harlan posted a link to this:

Lehman Excel snafu could cost Barclays dear.

The last time spreadsheets/Excel made the headlines was the SocGen debacle, I think,

Anyway in this particular case I would dispute that its some kind of Excel blunder. More appropriate:

  1. Unrealistic deadlines
  2. Inexperienced poorly supervised users
  3. Lack of testing and checking
  4. Software fail – the pdf converter didn’t respect the user view
  5. Other?

What do you think?

cheers

Simon

Why should Microsoft enhance Excel?

Wednesday, 15th October, 2008

Lots of people are bemoaning features, lack of features, or bugs in Excel. (not me of course ;-))

Anyway the thought struck me, why should they bother?

I’m not saying they have or they havent. I’m not saying they should or they shouldn’t.

I am asking what’s in it for Microsoft?

We’ll all end up updating eventually right?

How big an impact would it have on their finances?

Whats the cost benefit?

Would that money invested in developer effort be rewarded with

  • more sales?
  • faster take up?
  • sales at a higher price?
  • Reduce the rate of decline of MS Offices market share?
  • Reduce Vistas impact on OS market share free fall?
  • cross selling opportunities?
  • all of the above?
  • Others?

I’ve put some thoughts about their ribbon investment in the new traditional place.

I know kudos-wise if everyone on-line exclaimed wow Excel v n is mega super-de-duper fan-chuffiin-tastic, that would be great. But business-wise does it stack up?

And don’t be quoting any ‘quality’ or ‘excellence’ claptrap. Those are great for motivational speakers, Return On Investment is what keeps firms in business.

  • Where is the ROI on catching up with OOo in worksheet functions? (or UI),
  • Where is the ROI in updating VBA?
  • Where is the ROI in making SUMIFS as powerful as SUMPRODUCT?
  • Where is the ROI in providing multiple UIs to suit multiple uses/users?
  • Where is the ROI in fixing charting?

Personally I think many of these things would pay back, and I hope OpenOffice eating their lunch will encourage them to divert resources to these sorts of areas.

(I haven’t considered the opportunity cost of investing in these over some of their other technologies (search? Yahoo!?), feel free to do so in the comments)

What do you think? On an ROI basis what 3 things would give the best rate return? why?

Is it even worth investing in desktop apps? or do you think the cloud/unimind will solve all that?

let us know

cheers

Simon

CrossCheck XLM

Wednesday, 15th October, 2008

Every good beancounter knows you should always cross cast to check tables add up the same in both directions.

1 1 1 << sum
1 1 1 << sum
1 1 1 << sum
1 1 1 << sum
^^ sum ^^ sum ^^ sum ?? sum

The problem comes when you do that bottom right corner. Oh! sum from above or sum from the left? hmm.

Or add another sum below or to the side. Oh and if summing the horizontal and the vertical then you should really compare them and flag an error if they don’t agree. Thats starting to be a big formula.

Instead of SUM what about a function that can add both, compare them, return one if they match and return a text string if not? glad you asked, here is exactly that in glorious XLM – I have thousands of these in a live model on auto calc, that is still responsive. (needs to go in column D, crosscheck in row 1, otherwise that ‘D8’ is going to be looking in the wrong place, define name as function. etc)

CrossCheck
=ARGUMENT("Arg1",1)
=ARGUMENT("Arg2",1)
=ARGUMENT("ErrorMessage",2)

=RESULT(1+2)

=Arg1-Arg2

=IF(ABS(D8)<0.01)
=RETURN(Arg1)
=ELSE()
=RETURN(ErrorMessage & " : " &D8)
=END.IF()
=RETURN()

Of course I should have passed in that 0.01 materiality level, or set up a name for it, but I’ll leave that as an exercise for the reader.

Use it thusly:

=CrossCheck(Sum(A1:A4),Sum(A5:C5),”arrgghh!”)

Anyuse?

Any obvious blunders? suggestions?

cheers

Simon

OpenOffice 3.0 out

Monday, 13th October, 2008

OpenOffice.org overwhelmed by demand for version 3.0

I’m guessing thats a nice problem to have

Linux format has a less than overwhelming, and rather light review of the new version, eventually deciding it does just about justify the full increment

As a matter of interest – does/would anyone use that start centre mentioned in that article? I don’t/wouldn’t.

I’m looking forward to reviewing the download figures once the 3.0 effect has filtered through.

I think I’ll wait for the initial rush to ease off before I download it.

Anyone using 3.0 already? views?

cheers

Simon

[ps I love that VBA compatibility is mentioned as a big win – makes me think that the OOo folks ‘get it’]

Volatile UDFs

Sunday, 12th October, 2008

I had a discussion with a client recently about using volatile functions (User defined ones that is).

(volatile as in it calculate any time Excel calculates, not anything to do with our financial markets!)

My view is to avoid them so that Excels smart calculation can do its stuff, and so that cell auditing works.

Their view was pulling variables in directly from certain cells reduces the number of parameters and therefore simplifies the function.

I think both views are valid and it comes down to relative priorities. But the performance of the function has an impact, big slow VBA clonkers would steer me away from volatile, but if it was a low overhead thing, then maybe not so much of an issue.

Anyway it struck me that it would be worth knowing what the difference was between a volatile function and a non volatile one. Its about 200 per hour!

In my extremely limited testing a volatile function would get called 200 times more than a non volatile per hour.

I used this simple xll available here to just keep incrementing every time Excel calculated whilst I was building a model.

Its dead simple just:

long int lCount = 0;
EXPORT long int VolatileCalcCount ( void )
{
return ++lCount;
}

You could do the same in VBA, this is just less intrusive. Just add that as a worksheet function somewhere out of the way and watch it increment as you work.

Excel basically recalcs any time there is any change, so I must have done about 200 edits in the hour. Thats only 3 a minute – doesn’t sound like much – they must have been big ones ;-).

Let us know what score you get

cheers

Simon

VBA to the rescue

Saturday, 11th October, 2008

something for the weekend: Rob sent me this wtf the other day

Cheers

Simon

Visitor browsers

Friday, 10th October, 2008

Browser breakdown for codematic for Sept 08:

MSIE 60.20%
Msie 8.0 0.80%
Msie 7.0 23.90%
Msie 6.0 33.00%
Msie 5.5 2.20%
Msie 5.0 0.10%
Msie ? 0.00%
FIREFOX 27.20%
Firefox 3.0.3 1.40%
Firefox 3.0.2 1.50%
Firefox 3.0.1 16.70%
Firefox 3.0 0.70%
Firefox 2.0.0.7 1.70%
Firefox 2.0.0.4 0.10%
Firefox 2.0.0.17 0.80%
Firefox 2.0.0.16 2.50%
Firefox 2.0.0.14 0.10%
Firefox 2.0.0.12 0.10%
Firefox 2.0.0.1 0.10%
Firefox 1.0 0.10%
NETSCAPE 0.20%
Netscape 7.01 0.00%
Netscape 5.0 0.00%
Netscape ? 0.10%
Others 12.10%
Mozilla 3.60%
Unknown 3.50%
Safari 2.90%
Opera 1.40%
Microsoft Data Access Component Internet Publishing Provider Protocol Discovery 0.20%

Interesting bits:

  • MDAC – whats that doing there? WebDav?
  • 60:40 IE: non IE
  • IE 6 still the single most popular browser
  • Firefox users seem to upgrade faster than IE users
  • Aren’t there a lot of Firefox versions in use!

Summary IE6 : IE7 : Ff3 = 33 : 24 : 17

What do your stats look like?

cheers

Simon