Thursday, July 31, 2014

Feral systems

Over the past few days, I've been working on a supplement to my research proposal. I started by searching Google Scholar with the string "ERP management", looking for recent (post 2010) papers. This probably led me to an organisation with which I was not familiar - the International Conference on Enterprise Systems, Accounting and Logistics, aka ICESAL. There were some interesting papers to be found there, mainly about management accounting. This isn't really what I am looking for, but reading some of those papers led me to other papers and eventually I found gold.

I stumbled upon a paper entitled "Post ERP feral system and use of feral system as coping mechanism" [sic - yes, there are no articles used in the title] which upon reading seemed very similar to my research, at least on a theoretical level. On a practical level, it is completely different, which is actually a good thing. I could use the theoretical basis of this paper as justification for mine. The research on which the paper is based was conducted in Malaysia, with two large organisations using SAP, with the data being collected by interviews. One is allowed in doctoral research to take someone else's research and apply it in a different area: mine will be in Israel with many respondents using Priority, with the data being collected by questionnaires. 

Once I had this paper - and I wrote a long discussion about it - I was able to start researching 'feral systems' which led me to some very interesting papers, all of them published within the past few years. In some of them, I found arguments explaining why such systems are frowned upon - exactly what was required.

What is a feral system, I hear you ask. One paper which I quoted defines feral systems as being "the usage of information technology that deviates from the standard organisational norms and which exists beyond the control and/or knowledge of the organisational IT management". Perfect. This is actually a wider scope than I had envisaged,  but if the shoe fits....

The material which I present in the supplement is intended only to satisfy the research committee: I don't think that it will affect my hypotheses. On the other hand, I'll have plenty of literature to write about. I noticed that I was much more critical of the literature for this supplement; when I write "critical", I mean "able to discuss the paper by showing its good points and also what has not been mentioned". The literature review for the thesis has to be critical so it's good that I'm developing these skills now.

I have sent a draft of the supplement to my mentor but I don't expect a response for the next few days.

I wondered why I hadn't found any of this material for my original literature search, which in a sense lasted two years. The main reason, I think, is that once I had found the term "end user computing", I used it almost extensively in my searches, and didn't try looking for anything else. Also, I admit that my literature search was somewhat directed: I would search for a specific subject and once found, I would move on. There wasn't much room for serendipity.

This time, I tried a new search string which did indeed lead me to something different.

Tuesday, July 29, 2014

Research proposal accepted

Three weeks ago, I wrote about submitting research proposal for my doctorate. As I then wrote, the proposal will be given to two members of the research committee who will read it thoroughly and present it at the next research committee meeting. I spent the latter half of last week in anticipation of their decision; on Friday afternoon I was informed that my research proposal had been accepted . To quote my mentor: Fortunately, unlike many other DBA students, you do not have to resubmit the proposal for the next meeting on 23 September. However, you will need to address some specific points and respond to them in writing.  Yesterday I received the detailed review of the research proposal (completed by the two readers) and it makes very interesting reading.

One reader seems very much "on my side"; whilst he has criticisms, he generally accepts what I have proposed. His concluding remarks are In my opinion, this is a viable area of research and the student has considered it in great detail before writing the research proposal. After this statement, he then mentions that he had met me and that we had discussing research ideas - this gave the game away as I realised then who this reader was (a guest lecturer who came to the MBA course).

The other reader had different criticisms, the most important being If one accepts the topic the candidate has selected as being valid then a coherent argument for the research is made. However I do not think the candidate makes a convincing case. I cannot see that the topic, even if it were valid in principle, has the degree of business or management relevance expected in a DBA. He is also concerned that most of the literature appears to be relatively dated. Many of the key references are 5 years old or more. This seems odd in such a fast changing field.
 
My task now is to answer these criticisms. The major part will be finding convincing arguments in (preferably) recently published papers which make a case for why using EUC in an ERP environment is a bad idea. I have a strong suspicion that I won't be able to find such papers as I don't think that anyone in academia has addressed this problem. I referenced articles explaining this position but these are opinion pieces and not papers published in peer reviewed journals.

I will have to look for papers with completely different search parameters in order to find suitable material. I will also write to the consultant whose blog crystalised the research topic for me. I mentioned a few months ago a seminar which was to take place in Delft and wrote that there is a paper being presented which is very close to my area. I have been in contact with the person who gave this paper but he hasn't sent me a copy. He suggested a few weeks ago that I telephone him - maybe I'll do that this evening.

Thursday, July 24, 2014

Improving the In-basket

Four years ago, I wrote a series of blog entries about an exam which the Occupational Psychologist uses called 'The In-basket'. A few weeks ago, we had a discussion about this exam in which she outlined a few changes/improvements which she wished made. There is, unfortunately, no correlation between the apparent simplicity of each change and the simplicity of implementing each change: frequently, seemingly simple changes are hard to implement whereas seemingly complicated changes are easy to implement.

So it was this time. The hardest change to implement was probably the addition of tags (or labels) to each post; the initial idea was that the examinee would go over the items in the inbox, define a tag for each item, sort according to the tags then deal with each item. The implementation required several hours spread over a number of days before it was bug-free. The addition of such tags (and other new fields) required changes in the structure of the file which is written holding a log of all the examinee's action - and of course, the parser which later reads this file has to be changed and the program which prints out the results in human form also has to be changed.

As the exam itself is quite complicated and certainly not structured like a regular, multiple choice, questionnaire, it took some time to remember exactly how the program works and how the output is defined. Only then could I start adding new functionality. One of the side effects of my rewriting parts of the program is that the file structure is now documented.

Originally, the results program used to create its output via Word automation, laboriously sending line after line to the Word object which was created at the beginning of the routine. Apart from being slow, this approach also suffers from the problem that should the program fail during the output stage, an invocation of Word would be left orphaned in memory. These days, I create the output using HTML which is faster and safer, then insert the complete HTML code into Word. Sometimes the Word code can be converted very quickly into HTML, but in this case, it took about an hour and a half to rewrite the code, which is about ten times as long as it normally takes. 

I have now passed the program onto the OP for her final testing before we deploy the new versions. I hope that there won't be much to correct or improve, although now my renewed understanding of the program is much deeper than it was a week ago.

Tuesday, July 22, 2014

And now for something completely different

Our satellite film channel decided to hold a mini-Monty Python festival on Saturday, presumably in recognition of the comedy troupe returning to the stage after 40 years. They showed the following films
  • Eric the Viking
  • And now for something completely different
  • A fish called Wanda
  • A liar's autobiography
  • The meaning of life
Apart from "A liar's autobiography", I have seen all the others, although not for many years. I bought the book on which "Liar" is based many years ago.

I've seen "A fish called Wanda" several times and have always enjoyed it. I didn't bother watching or recording it on Saturday as I have previously recorded it.

Yesterday I watched "And now ...", which is a collage of sketches which were originally broadcast on the TV show. I'm sure that 40 years ago (1971-73), I found some of this material hilarious, but yesterday I was left totally cold. The last five minutes of the film featured some of the strongest (or most well-known) sketches - the dead parrot, the lumberjack's song and blackmail - but only the latter succeeded in raising the thinnest of smiles on my lips. The film simply wasn't funny.

Whilst it's possible that I have become what the film depicts, again in one of its final sketches, an accountant and thus per se extremely dull, I think that it's more that the material hasn't weathered well over the years. MP was a breath of fresh air at the time, with its surreal humour, general irreverence and snappy pace (cutting to a new sketch when the old one became tired, even before the punch line - or maybe there was no punch line), but now we're accustomed to these tricks. Which means that the material - as opposed to its presentation - wasn't particularly funny even then. 

I was never partial to the slapstick, such as the Ministry of Silly Walks (which thankfully didn't make the film) or the Upper Class Twit of the Year award. Maybe I was primed by reading an article in a newspaper's web site (unfortunately I can't find it now) which also didn't think much of their legacy.

Rice and beans

This is a recipe which I saw displayed on Jamie Oliver's tv show a few weeks ago.

Ingredients
1 can of mixed beans
2 cups of cooked rice
a pinch of cumin
fresh lemon juice

Empty the can of beans into a colander, then wash to remove excess sauce. Place them in a frying pan which has been pre-heated with a little olive oil. Fry the beans for about five minutes. Add a pinch of cumin at some stage. When the beans are dry and hot (in Jamie's programme, the skin of the beans cracks open), add the rice and a dash of lemon juice. Stop heating then mix well before serving.

I tried this out a few days ago and my experience was slightly different. Firstly, I couldn't find a can of mixed beans so I bough one can each of three different types of bean (black, white and brown (kidney)), mixed them together then used part of the resulting mixture (the rest is now stored in the fridge). I also didn't have any cooked rice handy, so first I had to cook some rice before I could start with the beans. Other than that - the dish is very easy to make and tastes very nice. I probably added too much lemon juice but that wasn't a fatal flaw.

Saturday, July 19, 2014

Statistics with SQL (Firebird)

Yesterday I was implementing a fairly advanced statistical analysis ("Testing a hypothesis about two population means when the combined sample size is at least 30", as my doctoral statistics text terms it), during which I realised that I could improve one phase of the calculation by changing my SQL queries. On the basis of that change, I decided to write an entry explaining some simple statistic measures and how to calculate them. I will use two approaches to the calculations, one naive and one (hopefully) sophisticated and measure the difference between the two.

The statistic measures are
  1. The number of data items (n)
  2. The mean (average) of the data items - the sum of the values divided by the number of items (m)
  3. The standard deviation of the data - this is a measure of how dispersed the data are from the mean (sd). This is obtained by initially calculating the difference between a datum and the mean, then squaring this number (as a result, all the values will be positive). These values are summed then divided by the number of data items; the result is called the variance. The standard deviation is the square root of the variance.
In normal distributions (frequently referred to as "the bell chart"), approximately 66% of the data will be between one standard deviation below the mean and one standard deviation above the mean. Approximately 95% of the data will be between two standard deviations below the mean and two standard deviations above the mean.

An everyday example of such a distribution is the IQ test. Unlike real distributions, the mean of the IQ is artificially fixed to be 100; the standard deviation is about 15. Thus 95% of the population have an IQ between 70 and 130.

Enough theory and onto the experiment. I am using as my data set a table which holds the results for all the people who have taken a certain exam. From one exam, several results are obtained, each for a particular scale. To simplify matters, I will take the results from only one scale. Thus the table structure is: pid (personal id), scale and num (test value), where pid and scale comprise the composite primary key.

The naive approach
We start with the simple data query select num from results where scale = 1. This will return one record per person. A simple loop will be sufficient to calculate both n and m:
var
 count, total: integer;
 mean, diff, sd: double;
 qResults: TSqlQuery;

begin
 count:= 0;
 total:= 0;
 with qResults do
  begin
   open;
   while not eof do
    begin
     inc (count);
     inc (total, fields[0].asinteger);
     next
    end;
I am deliberately not closing the query because it will be used again to calculate the standard deviation. At this stage, the variable count is n, and the mean is calculated by dividing total by count. Once the mean is known, then the standard deviation can be calculated in the following manner.
   mean:= total / count;
   diff:= 0.0;
   first; // iterate over the query again
   while not eof do
    begin
     diff:= diff + (fields[0].asinteger - mean) 
                 * (fields[0].asinteger - mean);
     next
    end;
   close
  end;
 sd:= sqrt (diff / mean)
end;     
When run against the database, the above code took 562 ticks on average. This can be divided into three sections: opening the query (15), calculating the mean (270) and calculating the standard deviation (277). This time is dependent on the size of the data set.

The sophisticated approach
Instead of using a simple query and calculating all the variables in code, SQL can be used to provide the calculations.

The number of data items can be calculated very simply with this query: select count (*) from results where scale = 1. The mean can be calculated in two ways: either the total is calculated (select sum (num) from results where scale = 1) and then divided by the number of data items, or it can be calculated directly (select avg (num) from results where scale = 1). Whichever approach is used, the two queries can be combined into one - select count (*), sum (num) from results where scale = 1.

Once the mean is known, the standard deviation can be calculated. This again can be achieved by using one query - select sum ((num - mean) * (num - mean)) from results where scale = 1, then dividing this by the number of values and then taking the square root. I would have liked to parameterise this query (ie passing mean as a parameter) in the form select ((num - :p1) * (num - :p1)) ... but unfortunately the syntax of Firebird doesn't seem to allow me to do this. So I have to use an underhand trick and dynamically build the select statement, including the mean as a magic number.

When run against the database, the above code takes 399 ticks on average. The run time of these queries should be much less dependent of the size of the data set. One benefit of running the queries on the SQL server is that there is much less network traffic - in fact, there is basically none in the sophisticated approach.

Following are the complete results from my tests. The timing is not particularly accurate as Windows is a multi-tasking environment and theoretically any event - such as either the anti-virus program or Dropbox waking up - would upset the timing. That's why I ran each test several times. The winner - by a clear margin - is the sophisticated approach, running in about 70% of the time of the naive approach (or 40% faster). I have to admit that this is less of a difference than I had expected.

Naive 1 2 3 4 average
Open query 15 16 15 15 15
calculate mean 282 265 250 282 270
calculate sd 297 281 266 266 277
Total 594 562 531 563 562

Sophisticated 1 2 3 4 average
calculate mean 187 188 188 172 184
calculate sd 219 203 203 234 215
Total 406 391 391 406 399

After I had run my tests, I decided to see what I could find on the Internet about calculating standard deviations. It turns out that Microsoft SQL server has two dedicated functions to do this (STDEV and STDEVP), the difference being that the first is used when calculating on a sample whereas the second is used when calculating on the entire population. There is a slight difference in the formula to calculate the standard deviation of a sample, but when the sample is large (I would say over 100), this difference can be ignored.

Firebird has no internal function to calculate the standard deviation. I read some discussion of trying to do this, but an elemental mistake was made: the code presented sums the difference between the value and mean instead of summing the square of the difference between the value and the mean. Unfortunately I have no idea when the above discussion took place and whether someone posted a reply, pointing out the error.

[SO: 3590; 2, 13, 35]
[MPP: 441; 0, 1, 6]

Thursday, July 17, 2014

EAST and research questionnaires

One website which I find very interesting is the Psy-Fi Blog, which as its banner states is a sideways Look at Psychology and Finance; in other words, it is concerned with behavioural economics. Unfortunately, the blog took a year long rest, but recently returned to life with a series which seems to be an A-Z of behavioural biases, most of which are very difficult for us to overcome.

One entry in the past few days which was not part of this series was this one, which discusses how using behavioral techniques [can] help improve citizens' responses to various government initiatives. This entry caught my attention, not because I'm particularly interested in improving responses to government initiatives, but because I'm interested in improving responses to research questionnaires.

True, I haven't got to the stage of distributing the questionnaires yet (when I will be very interested in obtaining a high response rate), but I did send out an initial consent letter to about twenty companies using Priority before I went on holiday and I have yet to receive a single reply.

The words in italics come from the Psy-Fi blog entry:
  • The "Easy" strand of the framework targets removing frictional costs; those apparently minor inconveniences that put people off whatever it is we want them to do - in my case, getting the companies to assent to participating in the research, then getting the respondents to reply
  • For the second strand, the "Attract" leg of the framework, the key is to personalize messages. In a complex and busy world you first need to get people to pay attention, and then you need to incentivise them to respond. It's not ethical to offer payment to people to complete research questionnaires; even if I did, I would be introducing bias as possibly only those who were interested in payment would return their questionnaires.
  • Pointing out that most people are compliant to some behavior will trigger our attraction to social norms. Maybe I will have to portray myself as someone who is desperate: 'without your small contribution, I won't be able to complete my doctorate and several years of study and research will go down the drain'.
  • Make it timely: the same intervention can have significantly different outcomes dependent on when it occurs. This really is irrelevant in my case, although again I could claim that I have to finish my research within three months otherwise ....
It is definitely important to encourage as many people as possible to respond to the questionnaire without introducing biases. In an ideal world, everyone to whom I send the questionnaire will respond, but somehow I can't see that happening.

Tuesday, July 15, 2014

Yehonatan (Jonathan) Geffen

When I emigrated to Israel in 1978, Yehonatan Geffen was the writer who in my eyes had replaced Amos Oz as the Israeli author to read. Geffen seemed to be multi-talented as he was publishing in at least three different fields, namely song lyrics, children's stories and a weekly column in the newspaper. I don't think that he had published any novels at the time, but he would shortly do so.

[Digression - a few stories about Amos Oz]
Oz though had only published novels, although they were very good; the most famous (at the time) being 'My Michael', which is about two students falling in love in pre-67 Jerusalem. As it happens, Oz was until 1986 a member of the kibbutz neighbouring mine; we used to use their swimming pool as at the time, my (old) kibbutz didn't have one.  It was maybe a week after I had emigrated, and I had gone with friends to use this pool; while I was resting at the deep end of the pool, I saw Amos Oz accompanying Jane Fonda.

Fast forwarding a few years, this neighbouring kibbutz hosted a get-together with my kibbutz, so my wife and I went over there (we presumably reciprocated at a later stage). We were shown a few houses, including one which was furnished in a style not normally seen in kibbutz houses (which tended to be small and furnished simply). My wife didn't catch on until we left the house, our hostess calling out 'Thank you, Amos' to the person whose house we had just visited.

A probably apocryphal story about Oz relates to the kibbutz accountant telling Oz that he is the most profitable branch in the kibbutz, and in order to increase his output (and thus profit), the kibbutz is assigning a few older members no longer fit for manual labour to help Oz. Someone obviously didn't learn enough economics.
[End of disgression - back to Yehonatan Geffen]

In 1978, Geffen had two records in current circulation. One was a satirical show called something like 'Small talk' (sichot salon), for which I didn't really have enough Hebrew at the time to appreciate. This show had been touring the country and the record was a recording of a show (I don't remember whether it was a live recording or a recording of the show being performed in a studio). The other record was a series of children's stories, some of which had been set to music: this was called 'The sixteenth sheep' and its musical director was Yoni Rechter, about whom I have written before. I didn't have so much of a problem understanding the lyrics, but I was much more interested in the music.

I have been given to understand that this was the first children's record which didn't talk down to its audience. The songs were performed by some of the top musicians in Israel at the time (Rechter, Yehudit Ravitz, Gidi Gov and David Broza) and were (mostly) written and arranged with great taste by Rechter. It is still as vibrant today as it was 35 years ago.

Here's someone else's take on the record, including a comment which I really love: The 'sixtieth [sic] sheep' is the best! I grew up listening to it, and so did my kids. The kindergarten song doesn't make you feel sad. In Hebrew its more like the kindergarten is in 'suspended animation', waiting for the kids to come back. The disc does have some *really* sad songs like the one about the beggar.

The record was turned into a musical show for which a few performances were held at the time. I remember taking the children from my kibbutz to Jerusalem in the spring of 1979 in order to see the show; I probably enjoyed it more than them. It seems like every ten years or so, a new production of the show is made: the cast may change, but the songs stay the same. 

At the time, I was involved with the cultural life of the kibbutz (mainly booking acts that I was interested in), so I jumped at the chance to book David Broza, who came in September 1979. After performing, he and his manageress (soon to be his wife) came to one of our rooms; we sat around, playing guitars and singing, and having a wonderful evening. At some stage, he signed the cover of my copy of 'The sixteenth lamb', making a pun on my name (which in English means something like delightful or pleasant - well, I did choose it for myself) and writing how delightful it was to play on the kibbutz.

In 1982, Geffen had three books published under the title "All my stains" (this was a Hebrew pun on the title "All my writings" which people frequently used): a book of song lyrics with music, a book of his collected newspaper columns and a book of various items from his satirical shows (including the skits from 'small talk' which I could now understand). I snapped this collection up.

Over the years, he has published novels, including one which I very much like, "The stone chair", which is a loose fictionalised autobiography, having sections which take place in Highgate, London. I continued to follow his life via his weekly columns (including a lengthy sequence about his life when he was in America for a few years) but it seemed that his performing days were over. Maybe the economics weren't suitable for the kind of show that he used to put on, maybe the climate had changed and people weren't prepared to listen to his satire anymore, maybe he wasn't in the country or maybe he was simply tired of performing (his columns over the past few years hint at medical problems; he is now in his sixties). Unfortunately, the newspaper which used to publish him nearly closed and in doing so, lost many of its reporters and writers, so we no longer have his weekly columns (and I had to switch to a new equivalent of a Sunday paper).

It was with great pleasure upon returning from holiday that I noticed a poster announcing a performance by Geffen on the kibbutz on Friday, 11 July. As this was a performance for children and not for adults, the show would start at 4pm. I instructed my wife to buy tickets. It transpires that Geffen is running a new version of the 'Sixteenth lamb' show along with new musicians who probably hadn't been born when the original record was released. The show itself was enjoyable (how could it not be with such good material) but the sound wasn't too clear, and the audience - mainly pre-school children - were rather loud.

One great advantage of going to a show on the kibbutz is that normally there is no problem in going backstage and meeting the artist(s). I had decided on a few items which I would take with me to get signed: the cover of 'The sixteenth lamb' (along with its Broza dedication), 'The stone chair' and one of the volumes of 'All my stains'. After the show, we went up onto the stage and asked one of the stagehand whether it would be possible to meet Mr Geffen. It would be no problem, we were told, but Mr Geffen is very tired and needs some time after the show to unwind. 

As we were at home (and I had the Sabbath evening meal in the slow cooker), we weren't in any rush and waited. At some stage, another stagehand told me that someone had brought their vinyl copy of 'The sixteenth lamb' to be signed; when Geffen wrote on the sleeve, the record cracked. I had already thought of this possibility and had prudently removed the record from the sleeve, leaving the former at home. 

Eventually Geffen appeared but he didn't seem to be very interested in us, despite my making all the sycophantic noises that one does in such situations. He momentarily perked up when I mentioned our mutual admiration for Randy Newman, but that was all I could get out of him. Obviously I wouldn't be successful in inviting him home for a cup of tea as he left the hall and went straight into a waiting taxi. Before doing so, he signed my items although it's hard to read his writing (I think that he wrote something like 'Thank you for the admiration, from all the herd, Y. Geffen' on the album cover). I belong to the generation that collected autographs, whereas these days, people collect selfies. Fortunately, the hall manager - a friend of ours - was present so he took the required photo. He hasn't sent it to me yet so I can't post it here.

I was disappointed at this outcome, but as my children pointed out a few hours later, I had no reason to expect more.


Monday, July 14, 2014

40 years ago/the lost summer

I returned to Britain from Israel around the 7th or 8th of July 1974. I can't date this exactly, but I do remember shopping for records a day or two after my return; one of the records that I bought - 'Like an old-fashioned waltz' by Sandy Denny - is marked as being purchased on 10 July.

I spent the first week after returning in a daze, moping around and doing very little (such as trying to wrap my head around Sandy's new record which was recorded in a style very different to The North Star Grassman). My parents had returned to Cardiff from Bristol the year before (before I had finished at school) and so we were living in our new but old house. It was a Victorian three floor building in which we had lived when I was a small child. Then (and presumably in the few months between Bristol and Israel), we had lived on the ground floor, but now, we were living on the second floor and it was all very strange.

After about a week, my friend Danny got in touch (he too lived in Cardiff) and told me about a weekend camp which was going to be held. These weekend camps were held about a week or so before the main Habonim summer camps and were intended to be the climax of the year for each local group; it enabled everyone to get into the right frame of mind before the national camps. These camps were more of a laugh than anything else, possibly a chance for the older members to remember what it was like to be an ordinary camper and not a leader.

We escorted by train a group of youngsters from Bristol to Rugby (I think) where we joined other groups. I was quite pleased to see these youngsters taking part in national activities as in the few years that I had been responsible for them in Bristol, they had been geographically challenged and had never had the chance to see how other Habonim groups had conducted themselves. Of course, once we arrived at our destination, I promptly dropped them and went about meeting people that I hadn't seen for a few years.

After this weekend, there was probably another week when I did nothing which has stuck in the memory before I set off to the national summer camps, this time as a cook. I don't remember where the camp was situated - probably in Staffordshire - but I do remember that it rained a lot during the first few days. I didn't have any experience in cooking and especially not for a hundred people under canvas, using only gas rings, but I quickly learnt. I soon discovered that I was unable to eat the food that I prepared, not because it wasn't tasty (it probably wasn't) but because after toiling on it for several hours, I couldn't stand the sight of it! Instead I was introduced to the delights of peanut butter, interspersing peanut butter sandwiches with omelettes.

When I was the age of the campers -12 and 13 - we carried our own bivouac tents when we went on a three day hike. But in the intervening years, the children had become softer and they simply walked between predefined sites where the tents (not bivouacs) had already been erected for them. Normally cooks didn't walk with the groups but this year we were trying out a different organisational structure (as I would put it now) and the cooks were part of the groups and indeed were responsible for buying and cooking food during the hike. This leads me to consider how and what we ate when I was a happy camper and I have absolutely no recollection of this.

The first day of the hike must have been a Sunday, for I remember being told that a shop in the village which was our destination would stay open especially for me until 4pm. At around 2pm, I saw that the group was moving too slowly for us to arrive on time, so I set off on my own in order to get to the shop before it closed. I did arrive in time, ordered whatever I thought was needed - for supper, breakfast and lunch the following day - then went off to find the campsite. Once I arrived and the food was delivered and stored, I laid down to rest - I fell asleep but awoke when I discovered that cows were inspecting our food store! The rest of the group must have arrived shortly after.

My 18th birthday fell on one of the days of the hike; I remember that there was a nearby pub, and the leaders in our group were taking turns to go to the pub and use its toilet facilities, as well as possibly having a small drink. I got my chance to celebrate my birthday along with a girl called Jenny ("Jenny, penny for your thoughts" - Peter Hammill, "Slender threads"), with whom I shared a now legal half pint of cider (although she was under-age).

The last day of the hike stays in my mind for some reason: we were camped in a flat park, and the children attached to me had decided that we would eat fish fingers and baked beans - something quite easy to cook on a gas ring. I set up one dixie (as we called the pots) on the gas, poured in a little oil, then started frying fish fingers. Obviously I couldn't cook enough at the same time for everybody so I had to cook in batches. I remember that the first batch took forever to cook (obviously the oil wasn't hot enough) whereas the final batch 'was cooked before I even put the fish in the dixie', as I used to tell it (the oil was now very hot).

When the camp finished, I returned to Cardiff for another month before starting my new life in London. In this period, I worked in a bakery for minimum wage; this served as my introduction to the food industry - after all, I was going to study food science. At first, I worked a shift of eight hours, but after a few days I was working twelve hours a day. I was capable of this, after my work experience in the kibbutz, but discovered that it left me with no life of my own. There were grown men with families working twelve hours a day, six days a week - obviously the pay must have been good, but what about quality of life?

I earned enough money to purchase a good quality stereo cassette deck which was equipped with both microphone and RCA jack sockets. This piece of equipment was to serve me for at least the next ten years as the basis of my 'recording studio'.

At some stage in mid-September, I packed up and moved to the Smoke, but that's another story in itself....

Thursday, July 10, 2014

Analysing Excel

Over the past few days, I've been revisiting some of the links which I had for my research but discarded. These are connected to people and companies who analyse spreadsheets and suggest how they can be improved. One person with whom I was in contact is Felienne Hermans, who is a researcher at Delft University of Technology; an offshoot from her PhD research is the company Infotron whose site allows one to upload an Excel spreadsheet and have it analysed for syntactic errors. Whilst sites like these are very interesting, they are not what I am primarily concerned with.

I decided to take a spreadsheet (belonging to someone else as mine are too simple) and analyse it at the Infotron site. Here are some of the results:

Sheet!CellFormulaRefactoring
Levels!J125 J124*3 Consider placing 3 in separate cell
Levels!J126 J124*6 Consider placing 6 in separate cell
Combined!F28 95298+91404 Consider placing 91404, 95298 in separate cells
Combined!F46 451583+218684 Consider placing 218684, 451583 in separate cells
Combined!F53 30306+59191 Consider placing 30306, 59191 in separate cells
International!E52 549788+9594 Consider placing 9594, 549788 in separate cells

What this report is saying is that one should never have a formula consisting of naked values (such as 3, 6 or 95298); instead, a formula should consists solely of cell references, like J124*J125. When one wants to know what J125 is, one goes to row 125 and sees what the row's name is. In programming, this is known as never using magic constants; for example, if one is writing a standard playing card game, then one should never use the naked value 52, but rather declare a constant "decksize = 52' then always refer to the constant decksize.

I discussed the report with the person who created the spreadsheet, who said that the spreadsheet was based on data whose source is Priority. I pointed out that there should be no problem as a worksheet will be created with (at least) two cells per row: description and value. When one has to add two separate values together, then one uses the cell references which hint at the cell's description. "And if I have to change a value or add something which doesn't feature in the original spreadsheet?", he asked, presumably referring to transactions which haven't yet been posted in Priority. The answer is just as simple: add another row to the spreadsheet, where the description is something like "Provision for rent (not yet posted)", and the value is whatever it is.

This activity overcomes the problem which I listed in my research proposal as referential integrity: it is not possible to know the source of a datum in the spreadsheet; it could come from ERP, it could be a personal projection, it may be invented or it may be the result of an earlier error. A malicious user could enter false data into a spreadsheet in order to support fraudulent activity.

As mentioned above, a tool such as Infotron's overcomes syntactic errors; it wouldn't detect a user adding a row with the description 'fraudulent activity' and the value '10000'! That, as they say, is left as an exercise for the reader.

Last week (June 2014), the Institute of Chartered Accountants in England and Wales (ICAEW) launched their Twenty principles for good spreadsheet practice. In this document the ICAEW recommend that organisations adopt a standard. The document contains several examples of good practice, but of course ignores the fact that maybe the program from which the data originated might have better data manipulation tools.

Tuesday, July 08, 2014

Research proposal submission

Yesterday evening I submitted my DBA research proposal; according to my mentor, the proposal will be given to two members of the research committee who will read it thoroughly and present it at the next research committee meeting which will take place in two weeks' time (apparently they meet once every seven weeks).

Depending on how one looks at this, this is either the first, second or fourth milestone in my doctoral candidacy (fourth if one counts each exam as a milestone, second if one counts all the exams as one milestone and first if one ignores the exams altogether). In a sense, I have been working on this proposal for over two years although the bulk of the work has been done since December 2013. The initial draft which I sent to my mentor in March is about 50% of the size of the proposal which I submitted yesterday.

The major problem when writing the proposal - and obviously the thesis - is that one is writing a document having no prior experience of writing such a document and with no real idea of the required level/depth. The mentor/supervisor knows the framework but not the contents whereas I know the contents but not the framework.

There are parts of the proposal which have hardly changed over the months: the introduction was the first section to be completed, followed by the literature review/background section. This latter part was about 80% fixed a few months ago, with minor changes constantly being made. The sections which have undergone the greatest amount of change are the hypotheses and the research method (including sampling and statistical analysis). In a sense, this is to be expected: I wrote the introduction and background on the basis of the literature searches which I have been conducting for the past few years; this is the material which I know well and is specific to my research. The hypotheses and statistics part are less specific (in other words, the structure, if not the content, is similar to that contained in other proposals) and less familiar to me, but more familiar to my mentor. Thus he had a larger say in editing this latter half of the proposal than he did with regard to the first half. It is very important to get the hypotheses correct (which is why we have spent so much time on them) as these are not allowed to change during the entire research period.

I wanted to ensure that the proposal is not rejected on a technicality: the (deliberately faulty) research proposals displayed in the IBR1 course all had a timetable, but I had not included one in my draft and  my mentor had not commented on this. I mentioned this in one of our most recent conversations; he said that he would check, and returned to me the same evening stating that a timetable was required. So I included a basic timetable.

Hopefully the research committee will accept the research proposal, allowing me to continue to the next stage. If they don't accept the proposal, they will give feedback on the areas which need to be improved and I will have five weeks in which to improve the proposal to the required level (the committee meets once every seven weeks but one has to submit the proposal two weeks before the committee meeting). One is allowed to submit the proposal a maximum of three times - as they say in baseball, three strikes and you're out. I'm hoping for no strikes.

The mentor is very pleased with the proposal; he frequently comments that it is very interesting and original. The latter word is a two-edged sword: original in this context means that there is no prior research in this area. This means that one hand, I am performing what might be seen in the future as pioneering research, whereas on the other hand, I have only a small literature base. I think that the research is also original because I have the advantage of working both in a corporate ERP environment and also in a small company which administers psychological tests. Had someone else performed research in this area, they might not have had the psychological knowledge which I have.

Once the proposal has been accepted, the next stage is the complete literature review which will appear in the doctoral thesis. I am unsure what I am supposed to do as I have already scoured the literature, looking for anything relevant. This is the problem which I referred to in the previous paragraph; I have literature in four supporting areas but none in the specific area which I am researching. Obviously, I will be dependent on my supervisor (the job description changes, but it is likely to be the same person) for initial guidance here.


Saturday, July 05, 2014

Swimming

I went swimming this morning. They say that swimming is like riding a bicycle: once one knows how to do it, one never forgets. I don't know when I last rode a bicycle, neither do I remember when I last went swimming (I definitely remember 2007 in Santorini, but I may have swum since). But it's true: once I hit the water, I could swim! Even my timing (breathing, etc) was fine.

I'm rarely at home when the pool is open (save weekends), and when the opening hours are suitable, it's generally far too hot. So I have to pass on this form of exercise. A few days ago, I read that the pool will be open on Saturday mornings at 8:30 am for 'health swimming' (adults who swim lengths), so I bought a new pair of goggles (which seem to be too small) and waited for Saturday.

I was the second person in the pool, but it rapidly filled up - there were about 15 people swimming when I left, and that's enough if everyone is swimming lengths. The water was the perfect temperature for swimming - not too cold, not too hot - and once in, I confidently struck out for the deep end, swimming breast stroke. The first length went easy, but about a third of the way through the return length, I realised that I couldn't keep up the same pace. I finished the length then rested for a few minutes. 

I managed to swim six lengths interspersed with breaks whose length increased every time. After I had finished - when I had no strength left - I got out of the pool and sat by the side, breathing heavily for about ten minutes. I'm not used to this strenuous exercise, although I imagine that if I go every week, it will be easier.

It's now two hours since I finished swimming: my breathing has long returned to normal, but I can feel the results in my arms and legs.

I should point out that when I was young, I used to swim frequently, at least once a week. When my children were younger, I used to take them to the pool a few times a week, and of course I would take the opportunity to swim. But they've grown up, I've got a dog (with whom I walk) and I've had a history of BCC - so I don't swim anymore. The early opening hour on Saturday now gives me a chance to swim regularly again.

Wednesday, July 02, 2014

Once

Nearly four years ago, I was writing about the film 'Once' starring Glen Hansard and Marketa Irglova; the paragraph concluded maybe one day it will be screened here. Well, Yes3 (the obscure/French/arty film channel) broadcast 'Once' the other day - in the early afternoon when most people would have missed it. I only caught it by chance by perusing the listings for the upcoming week.

This time around, I had the benefits of a good picture and subtitles, so finally I could understand all of what the characters were saying (although I note that sometimes Hansard's thick Dublin accent defeated the translators and they also didn't translate the few words of Czech). Watching the film on tv cleared up a few misunderstandings - for example, I had thought that 'the guy' knew the musicians who help him record; it seems as if he comes upon them on the street as strangers. I was also impressed by the strength of character displayed by 'the girl' - she persuades the bank manager to authorise a loan and negotiates a deep discount with the studio manager.

Somewhat surprisingly, IMDB displays an 8.0 rating for the film. Whilst I enjoyed watching it, I didn't think that it was that good. It's very natural, so there aren't any one liners to be picked up. The relationship between the guy and girl is platonic (although he at one stage wanted more until being turned down neatly) so there are no romantic scenes. This naturalism is what causes people to think that it was a true story - exacerbated by the fact that the actors did become a couple for a short while after the film. Let us not forget that the film ends with the girl reunited with her husband and the guy flying to London to hunt down his ex-girlfriend. So the guy did not get the girl.

I was a bit disappointed by the music: two songs ("Falling slowly" and "When your mind's made up") are repeated, as opposed to using other songs.

Tuesday, July 01, 2014

Crossroads

About a year ago, I wrote a song called 'Crossroads', the first verse of which is

Paused at the crossroads
Which way should I choose?
One way’s a winner
The other way I lose

Paused at the crossroads
Turn left or turn right?
One way is sunshine
The other way is night

I have just discovered that there is a bluegrass song by Bill Monroe called 'The old crossroads', whose lyrics begin

The old crossroads now is waiting
Which one are you going to take
One leads down to destruction
The other to the pearly gate

The similarity is intriguing. I wonder now if I had considered writing 
Paused at the crossroads
Which path should I take?
Maybe I did think of this line but then had problems thinking of a suitable rhyme with "take". "Pearly gate[s]" is not a phrase which I ever use so this wouldn't have been in my mind.

Of course, harmonically, my song has absolutely nothing in common with bluegrass.