New place for this blog

More recent posts can be found at http://blog.dominicsayers.com

Why the iPhone is like Internet Explorer 6

I mentioned to a friend that I thought the iPhone presented the same dangers as Internet Explorer 6 did when it was released. I sense that this point is not as obvious as it seems to me, so here goes with a brief explanation of what I mean.

Internet Explorer 6 was a major advance on its predecessors as an application development platform. What’s more it was effectively ubiquitous – Microsoft gave it away with Windows and everybody had Windows.

In the corporate environment, this meant that IT departments could write client applications for the browser instead of using Visual Basic or worse, Java. These browser-based applications needed no roll-out programme and no desktop provisioning – both expensive headaches.

Whoopee! said corporate IT departments, and rushed to develop applications for the new platform.

Roll the clock forward to today and the folly of that approach is clear. Corporates are now stuck using those same client applications because either they can’t afford to develop new ones, or they bought a third-party application that is still business-critical, or the developers have left the company and they simply dare not touch the code. And those applications only work on Internet Explorer 6 because it turned out it wasn’t a de facto standard after all, it was a proprietary cul de sac.

So I sit every day at a desk in a bank, trying to use Internet Explorer 6 in a world where most people have given up trying to retain compatibility with this out-dated platform. And around the world there are millions like me whose employers or clients are saddled with this dinosaur because of an expensive mistake made years ago.

Why do I believe we are in danger of making the same mistake again?

The number of applications developed for the iPhone is astonishing. People develop for this platform because it is better than its predecessors. And although it is not quite ubiquitous, it certainly has a market share that makes the potential audience for an iPhone application very attractive indeed.

This has not escaped the notice of corporate IT departments. I imagine client applications for the company CRM system are being developed right now for the mobile salesforce of thousands of companies.

And so history repeats itself. The iPhone is a proprietary platform whose future is not yet clear. It could set a de facto standard for future mobile platforms. Maybe. I think it more likely that Android and Windows Phone 7 will go their own way. And corporates will have bought into applications for a platform that will go away before the business requirement does.

By all means buy or write applications for the iPhone, but please factor in the cost of replacing those applications in a few years. I don’t want to be carrying round a five-year-old iPhone in 2015 just because my company depends on an application that only runs on that old thing.

Graeme Swann

Shakib Al Hasan‘s century in the recent test match against New Zealand means that Graeme Swann is now the only member of my all-rounders table without a test match hundred to his name.

This brings into question whether Swann’s presence in the table is due to his being a genuine world-class all-rounder or, on the other hand, a statistical anomaly.

In my first post I talked about what makes an all-rounder. The underlying principle is that the player should merit selection as either a batsman or a bowler. If he is unable to do one or the other for a given match, is he still worth picking?

On this basis I think calling Swann an all-rounder is a bit of a stretch. If he couldn’t bowl he wouldn’t play. His usual batting position is 9, which he may feel is unfair but reflects the consistent opinion of a number of captains and coaches since 2008.

So why is he in the table? My batting criterion is an average over 30. If I raised this to (say) 35 I would be excluding Ian Botham, Shaun Pollock, Andrew Flintoff, Kapil Dev and Chris Cairns. All world class all-rounders according to conventional wisdom.

His average reflects the number of not outs in his 21 test innings. Here is how his average has changed since the start of his career:

Graeme Swann's batting average

I think it will settle in the low- to mid-thirties for the remainder of his career. The peak of 45 was a result of two significant not-out innings in 2009 (63* against West Indies and 47* against Australia) and I don’t think he’ll approach those heights again. But if he keeps above 30 he’ll stay in my table.

So are the criteria wrong? How can I exclude a player like Swann when his figures bear comparison with the best of all time? All I can think of is  excluding players who bat at 9, 10 or 11 in the batting order. This seems even more arbitrary than my existing criteria and I don’t like it much.

Any ideas?

Shakib Al Hasan

After my recent post about the World’s Best All-Rounders a reader asked me why I hadn’t included Shakib Al Hasan of Bangladesh in my list. The reason was simple: he had not qualified. At the time he lacked 2 wickets for the 50 wicket qualification, and his batting average was 29.79, meaning he just missed out on the batting average qualification too.

Since then he has played 2 more tests against India and is competing in a third against New Zealand. His bowling continues to impress and he has another 10 wickets to his name since I last wrote, meaning he now easily qualifies for his bowling both on aggregate and average. Not bad for a 22-year-old.

His batting, however, wasn’t doing him any favours as an all-rounder with scores of 17, 17, 34 and 7 in the games against India. All this changed yesterday when he enjoyed a whirlwind partnership of 145 with Mahmudullah and was unlucky to be dismissed for 87, bringing his test average up to 30.24. As I write, then, he is one of the World’s Best All-Rounders.

He is joined as a debutant in the table by Graeme Swann, who has also qualified since my last post:

Player	Matches	Runs	Bat Av	Wkts	Bowl Av	Ct	X-factor
GS Sobers (WI)	93	8032	57.78	235	34.03	109	3370
JH Kallis (ICC/SA)	137	10843	54.76	261	31.55	155	3192
AW Greig (Eng)	58	3599	40.43	141	32.20	87	1985
TL Goddard (SA)	41	2516	34.46	123	26.22	48	1550
BM McMillan (SA)	38	1968	39.36	75	33.82	49	1359
KR Miller (Aus)	55	2958	36.97	170	22.97	38	1313
IT Botham (Eng)	102	5200	33.54	383	28.40	120	1193
Mushtaq Mohammad (Pak)	57	3643	39.17	79	29.22	42	1102
Asif Iqbal (Pak)	58	3575	38.85	53	28.33	36	963
SM Pollock (SA)	108	3781	32.31	421	23.11	72	947
ER Dexter (Eng)	62	4502	47.89	66	34.93	29	840
ST Jayasuriya (SL)	110	6973	40.07	98	34.34	78	761
Imran Khan (Pak)	88	3807	37.69	362	22.81	28	633
GE Gomez (WI)	29	1243	30.31	58	27.41	18	490
Graeme Swann (Eng)	16	525	32.81	69	30.69	10	445
JDP Oram (NZ)	33	1780	36.32	60	33.05	15	376
JR Reid (NZ)	58	3428	33.28	85	33.35	43	374
N Kapil Dev (India)	131	5248	31.05	434	29.64	64	313
MH Mankad (India)	44	2109	31.47	162	32.32	33	311
A Flintoff (Eng/ICC)	79	3845	31.77	226	32.78	52	263
Shakib Al Hasan (Ban)	17	877	30.24	58	29.96	8	248
CL Cairns (NZ)	62	3320	33.53	218	29.40	14	206
IK Pathan (India)	29	1105	31.57	100	32.26	8	119
DL Vettori (ICC/NZ)	98	3789	30.55	316	33.56	55	112

(I’ve omitted the pre-war players this time)

Dan Vettori has slipped down a place. There are now four current players in the table. I don’t know of anybody else who’s near to qualifying except perhaps Stuart Broad. Any suggestions?

A reader writes

Just got this interesting email. My reply follows.

Dominic,

I’m sorry if this email is, in any way (from the subject down to the signature) offensive, bothering, time wasting or anything like that to you. The sole purpose of this email is to better understand some things about your kind of career/profession, which is also mine. Furthermore, I’m not asking for anything besides information — and from your site I infer that you like to share your thoughts, so…

First, let me introduce myself. I’m a consultant working for a multi-national company. My specialty is System Architecture, and while the title is vague, what I actually do is design systems through the use of UML diagrams, a lot of documenting, etc. Usually I spend even more time implementing stuff, from C/C++ to bash shells to C# to COBOL even. I’m very fond of technology, and I chose my career when I was very young — about 10 years old. Back then my desire was to work with videogames (like most children with such love for gadgets and computers, I guess) but since I live in Brazil, I was forced to abandon that path some years ago.

I stumbled upon your website when looking for the maximum theoretical limit for the length of an email address. I needed that piece of information to define an entity in my database, and I wanted a number that was right from theory. When reading the article on the site, however, something struck me as odd. Your profile, on the left portion of the page, read “IT director”. But I was reading a purely technical article! That’s incredible. Here, both at my company and, as far as I can tell, my country, techies doesn’t go too far up companies hierarchies. So I decided that you were not, in fact, what I call “techie”. You are probably more, a guy with both systems expertise and managing skills. Still, such a profile doesn’t make much sense to me.

In my company, after as little as 5 years, you can get a managing position. And when that happens, it’s no longer expected that the worker learns how to implement specific things. In fact, the only thing people seem to be able to do afterwards is learn what a particular technology is capable of. For example, managers here know about SOA, about .Net, about ESBs, about Clouds. They know what they are and how they can help business. However, they don’t know how those stuff works. You, however, seems to know deeply at least about PHP, a good deal about REGEXes, etc. These skills, if not used, are forgotten, so I’m guessing you have a tight connection with technology. I’m sure you don’t look at source codes or solve complex math problems for your analysts, but I think you get in touch with them and probably study deep technology during breaks.

I assumed a lot of stuff up there but, if I’m right about most of it, my question is, how do you do it? How do you balance being a good Technologist with being a manager? Is there some kind of synergy between them? Does your bosses took your tech skills into knowledge as you progressed through your career? Or maybe I’m completly wrong and you don’t manage people, being just the big fish when it comes to technology (highly improbable, but still…)…?

I’m know that, from what I’ve written there’s a good chance that you take me for a fool. These questions have been storming my mind for some months now, and I’ve always wanted to ask someone with more time in the market than myself…

Thanks a lot for reading,

X

Dear X,

I was lucky enough to work in an IT department that valued technical expertise very highly. Even though I ended up as a fairly typical manager (death by PowerPoint) I was still in a community that respected technical skills, especially when they were married to communication skills that did not exclude non-technical people.

Unfortunately all good things come to an end and I am now working solo as a consultant in the Financial Service industry (I will update my web page soon). This work is somewhat stochastic and in my spare time I can indulge my technical skills (not as much as I’d like to though).

I don’t see any reason why you can’t be both a manager and a technician. But to succeed in the competitive world of corporate management you need a certain set of skills (sharp elbows and political nous). It’s a rare person who has both these skills and deep technical skills. This might be because deep technical skills used to come with poor personal hygiene and an introspective personality – I don’t know why, I’m not a psychologist.

Increasingly, technical skills are respected in the real world outside IT. I believe this will lead to “normal” people getting interested in technology (it’s already happening). You will see more managers (and company directors) with a technical background in the future.

Corporations value management skills more highly than technical skills. If you are in a management position then any technical activity is likely to be regarded as a waste of your valuable time. I agree with you this is wrong. I think it’s because today’s managers have no technical background and are suspicious of those who do. Thus, this will also change as more managers with technical skills are promoted.

Keep going with your technical activities. You are clearly a good communicator, so there’s no reason why you shouldn’t do both.

Just some random thoughts,

Regards,

Dominic

World’s best all-rounders

The traditional dilemma with a cricket team is how to get a guinea’s worth of value from a £1 budget. Ideally you’d want seven batsmen and five bowlers. That’s twelve players in an 11-a-side sport. The extra shilling usually comes from a talented individual who can contribute with both bat and ball.

The qualities this special player needs are to score as many runs as a specialist middle-order batsman and also to bowl as effectively as the other four members of the bowling unit. Anything short of this and you have not an all-rounder but an nearly-rounder. A squarer peg for the all-rounder hole.

Who are the players who have filled this role best in test match history? Names like Garry Sobers, Imran Khan, Ian Botham and Keith Miller spring to mind. But how to determine their effectiveness in filling that pivotal role?

Cricinfo’s statistics can give us the raw data. I now have a spreadsheet with all 2574 test match players and their raw figures. How can I use this data to identify the best all-rounders?

Out of 2574 players, 82 never batted long enough to get dismissed and 1053 never took a wicket. This includes 26 players who did neither in their test match career. No wickets and no completed innings: no hopers? Well Stuart Law is one of them so you decide.  Nevertheless, I think we can rule them out as all-rounders straight away. That leaves us with 1466 players who have taken a test match wicket and have a test batting average. Who were the best?

Let’s look at batting averages and bowling averages. You’d expect a test match lower-middle-order batsman to be averaging over 30. You’d also expect them to be taking wickets at less than 35 runs apiece. This gives us 120 all-rounders by this very broad and somewhat arbitrary definition. However the list includes players like Mark Boucher who played as a wicketkeeper-batsman but once took a wicket as an eighth-change bowler against the West Indies as a dull test match petered out into a draw.

So we need some qualifying numbers of runs and wickets to weed out these statistical anomalies. Again I will introduce an arbitrary cut-off of 500 runs and 50 wickets. This leaves us with 29 players who might be selected on the basis of either their batting or bowling alone.

We need a way of ranking these players as all-rounders. We can assess their contribution with the bat and ball by seeing by how much they exceed the qualifying criteria. In other words, how much higher is their batting average than 30. How much lower is their bowling average than 35? But there’s another factor – catching. Catches, as we all know, win matches. As your all-rounder you want a proper batsman, a proper bowler and a fielder who can pluck swallows from the sky.

By the magic of arbitrariness I have assigned an x-factor to each all-rounder. It’s simply a multiplier of their batting average in excess of 30, their bowling average below 35 and their number of catches per 100 matches. And the result:

Player	Matches	Runs	Bat Av	Wkts	Bowl Av	Ct	X-factor
GS Sobers (WI)	93	8032	57.78	235	34.03	109	3370
JH Kallis (ICC/SA)	134	10587	54.85	258	31.33	150	3193
AW Greig (Eng)	58	3599	40.43	141	32.2	87	1985
JM Gregory (Aus)	24	1146	36.96	85	31.15	37	1667
TL Goddard (SA)	41	2516	34.46	123	26.22	48	1550
GA Faulkner (SA)	25	1754	40.79	82	26.58	20	1537
BM McMillan (SA)	38	1968	39.36	75	33.82	49	1359
KR Miller (Aus)	55	2958	36.97	170	22.97	38	1313
IT Botham (Eng)	102	5200	33.54	383	28.4	120	1193
Mushtaq Mohammad (Pak)	57	3643	39.17	79	29.22	42	1102
Asif Iqbal (Pak)	58	3575	38.85	53	28.33	36	963
SM Pollock (SA)	108	3781	32.31	421	23.11	72	947
C Kelleway (Aus)	26	1422	37.42	52	32.36	24	929
WW Armstrong (Aus)	50	2863	38.68	87	33.59	44	888
W Rhodes (Eng)	58	2325	30.19	127	26.96	60	851
ER Dexter (Eng)	62	4502	47.89	66	34.93	29	840
ST Jayasuriya (SL)	110	6973	40.07	98	34.34	78	761
FE Woolley (Eng)	64	3283	36.07	83	33.91	64	716
MA Noble (Aus)	42	1997	30.25	121	25	26	635
Imran Khan (Pak)	88	3807	37.69	362	22.81	28	633
GE Gomez (WI)	29	1243	30.31	58	27.41	18	490
JDP Oram (NZ)	33	1780	36.32	60	33.05	15	376
JR Reid (NZ)	58	3428	33.28	85	33.35	43	374
N Kapil Dev (India)	131	5248	31.05	434	29.64	64	313
MH Mankad (India)	44	2109	31.47	162	32.32	33	311
A Flintoff (Eng/ICC)	79	3845	31.77	226	32.78	52	263
CL Cairns (NZ)	62	3320	33.53	218	29.4	14	206
DL Vettori (ICC/NZ)	97	3779	30.72	313	33.61	55	120
IK Pathan (India)	29	1105	31.57	100	32.26	8	119

Fascinating.

Garry Sobers is top, as you might expect. But the evil Jacques Kallis joins him in a top two who are miles ahead of the pack. More than 1000 x-factor points separate them from their nearest rival.

But this list looks good. It’s got all the usual suspects. Andrew Flintoff sneaks in although history will probably be less kind to him than our memories suggest. History’s judgement on Flintoff: a bit better than Daniel Vettori (but not as good a captain).

Notable absences: Richard Hadlee and Trevor Bailey averaged under 30 with the bat. Steve Waugh and Lance Klusener over 35 with the ball. Ravi Shastri averaged over 40 with the ball.

Graeme Swann will join the list when he has another 37 test runs (batting average 35.61, bowling average 29.65). Stuart Broad needs to get his batting average of 28.71 up a bit.

Note to self

No really. Move along there’s nothing to see here.

1. Herb Simon

2. Complex adaptive systems (esp. Nearly Decomposable systems)

3. Netflix mission statement

There’s clearly some academic work supporting the “we hold these truths to be self-evident” tone of the Netflix mission statement slides. Do some reading. Do some writing. This stuff matters.

NB Remember to credit Sean for the Prezi.