Evening ferry

As the weekend was free of children and grand-children (coincidentally, both of my children were in the north of Israel), I had plenty of time to complete a song which I had begun about a month ago. I am writing the following for myself whilst the memories are still fresh in my mind - I doubt that the nitty gritty details will be of much interest to anyone else.

Most of the music was written in the middle of July when I sat down at the piano and started playing something that sounded like a lullaby. I had been reading a music analysis of Nick Drake's songs, paying attention to the fact that the melodies in several of his songs start on the third beat of a 4/4 bar; this isn't an anacrusis (starting the tune a beat or half a beat before each bar). The original idea had an eight bar sequence in G which was first repeated then played three semitones higher (in Bb), followed by a final repetition in G. I sequenced this initial idea and played with it a bit, eventually realising that it was quite boring and would be better if there were one sequence in G followed by one in Bb. Further tinkering with the music led to an introduction which includes the sequence G F9 Em Cm7b5 - the final chord not being one that I normally use.

Shamelessly I pinched a few arrangement ideas from an alternative version of the arrangement which I created for my wife's masked song, but eventually the arrangement became autonomous. I also played around with the drumming, starting off with what sounds like a heart beat, then half time then normal time. A few evenings ago, I decided to change the instruments playing the introduction and coda, replacing them with real live electric guitar.

The words as usual were problematic. For a few weeks, there were no words or direction at all; one evening whilst walking the dog I hit on the idea of trying to convey the feeling that I had whilst sitting in a cafe in Gavrio after the last ferry had left for the day. I remember the same sort of feeling when we stayed at Yosemite national park in America before the birth of this blog: those that had come for the day had gone home, leaving the beautiful area to be explored by those spending the night there. It's an extension of something which we've been doing since that trip to America in 2005: instead of being day tourists, we've tried to travel to one location and become 'locals', seeing things that the day tripper wouldn't normally see.

Over a few days last week, I managed to write three enigmatic verses, but as usual I was lacking words for the bridge section. These came whilst showering on Friday night and getting into bed: every few minutes I would get up, race to the computer then add a couplet.

Yesterday I devoted my time to recording the song. After swimming, my first task was to play the guitar parts; this was fairly easy and only a few takes were required. Singing the song as usual wasn't quite so easy but again, not too many takes were required. I spent some time creating a harmony for the bridge from which I developed a choral sound, but this seemed inappropriate, although later on I would use a similar idea for the second bridge. Mixing, surprisingly, went very quickly; I remember doing this about ten times for an earlier song, trying to achieve an appropriate volume for my vocals with respect to the music, but yesterday the final mix was either the second or third attempt.

Time sheets and perceived value

As this blog entry is connected to my thesis topic, I think it only fair that I start with news about the final stage in my doctoral studies - the viva exam. It would be more accurate to write 'start with the lack of news' - the administrator of the doctoral programme informs me that the external examiners are very busy and that he has yet to fix a date for the viva. Onto a case in my company which I would have liked to include within the case studies presented in the thesis - although it's too late for that. Maybe I can bring it up in the viva as an example of how not to deploy an enhancement.

The CEO decided a few months ago that he would like certain departments to keep time sheets, showing on what specifically they have been working on (i.e. the number of a project, or the number of an engineering change) as well as including more general information, such as participation in meetings with customers or training. One can understand his point of view: it might well be that we are spending more hours on these activities than we are budgeting for, or to paraphrase Peter Drucker, one can't manage something without measuring it.

The task was passed on to the CFO, my manager; she along with the network administrator and I looked for programs which manage time sheets. Specifically we were looking at the sort of program that lawyers, accountants or architects would use to record billable hours. My first idea, that the external programs in use could manage time, was a dead end. Whilst Microsoft Word keeps a record of hours during which a document is open, the engineering programs that we use do not have this functionality. We found a few programs with which I played: these programs led me to believe that there is a continuum of approaches. At one end, there are programs that are geared to recording working hours (when a person clocks on or off) and that the task management was bolted on afterwards. The other end has a very programmatic database approach which was much more to my liking, but also twice as expensive as the work hours program (no names). I had also considered creating a module in Priority to handle the time sheets, and seeing the database approach strengthened my position.

I could already see that whilst the work hours program was easy for the end user (there are interfaces for the web and for mobile phones), it would be problematic to create the initial tasks. The CEO presented several requests which could not really be achieved using this program; they could be approximated but not really to the depth that he wanted. The database program had no external interface: it's a program that has to be installed on each user's computer.

Even before I had started work on a specification for a Priority based program, it was clear that there were certain advantages in using Priority: it's a known technology, it's flexible, it's already installed and it's (essentially) free - there would be no marginal cost in using a new module, provided that I wrote it. A further advantage is that several of the tasks could be created automatically as a result of events happening within Priority, such as the opening of a new project. There are also disadvantages, which basically boil down to the requirement to having Priority open at all times in order to have automatic time recording (i.e. Priority knows and can record the current hour when a button is pressed); this requirement would not be met for one group of users.

After receiving the go ahead to prepare a demonstration program, I spent several hours over one weekend working furiously on the module. Every now and then I would take the dog for a walk, during which I managed to solve problems that had arisen as well as considering new ideas. I showed the module (via Teams) to the CEO and CFO, from which more requests arose, which I solved over the following days. It would be accurate to say that there is a lack of fit between the interface required and the paradigm of Priority forms; I would say that I managed to obtain about 95% fit.

Then we showed the module to representatives of the two groups who would initially be using it: the engineers and the planners who interface between the salespeople and customers on one side and the engineers and production on the other side. I had predicted in advance that their reaction would be negative and I was not disappointed. Whilst they understood what the CEO stood to gain from the module, from their point of view it was 100% overhead which they could do without. Even after I explained that most of the tasks would be opened automatically and that the user interface was extremely simple, they were still adamant in their opposition.

As a test, I defined a few tasks for the CEO who barely uses Priority; after five minutes of training, he was able to report that the module was extremely easy to use. Of course, the ability to extract data in any form conceivable goes without saying. I have yet to hear anything from the other users who were supposed to test the module.

Why was I not surprised that the users would be against using the module? Because from their point of view, the perceived value of the module is not even zero, it's negative. Quoting my thesis, the Technology Acceptance Model (TAM) is based on four components, all based on the individual: perceived usefulness, perceived ease of use, the individual's evaluative judgment of the IT system and the individual's motivation to use the system (Davis, 1989). I can easily imagine that at least the first two components of the TAM would receive negative scores; some of the users (especially their managers) are strong users of Priority so at least the third component is positive. I imagine that their reaction would be even stronger if we tried using an external program.

Quoting the abstract of the thesis, "the research shows that companies that develop a tested framework for deploying enhancements are able to reap the promised benefits and so improve their business performance. The research also shows that user resistance is a major problem that has to be overcome in order to achieve a successful deployment". We managed to complete the first two stages of the seven stage framework, viz. the identification of a misfit (stage 1) and the preparation of a technical specification that defines the advantages that would accrue from the enhancement, the disadvantages and the financial cost (stage 2). We floundered on the third stage - obtain the signed agreement of all stakeholders on the specification, involving senior management wherever necessary.

I haven't heard any more on the topic of time sheets for the past few weeks, probably because almost all of the participants in the last Teams meeting on this subject are also involved to some extent with a much more important project, including the reception and installation of new equipment in one of the factories, the problems with getting technicians from Italy to install the equipment, and software interfaces (including a program that I wrote that receives a file from a design program then builds appropriate parts and a BOM based on the data).

A final quote from my thesis: there is no attempt to assess the economic validity or benefit of specific enhancements; there is no comparison between enhancements that provide economic value and those that merely satisfy whims. Does this proposed enhancement really provide economic value or is it a whim of the CEO?

New CPAP machine (part 2)

After nearly a week with the new machine - I uploaded the statistics file to the manufacturer's website in order to see the statistics. As I noted previously, the data is presented in the form of histograms, but clicking on a bar will give the numerical data. I am very pleased to see that most nights there have been four or less apneic events per hour, which is very good. This is probably more to do with the new mask than the new machine.

I developed a very simple spreadsheet into which I added the data: this is a bit awkward as the web site shows data per type of event (hours of sleep, apneas, leaks), what might be termed a vertical presentation, whereas the spreadsheet and my CPAP data program require a horizontal presentation, i.e. all the data for one date together. It took only a few minutes to copy the data into the spreadsheet and a bit longer to write a new module in my CPAP program to read the spreadsheet. Now that I have data, I will try to read the statistics file directly in order to discern any structure.

As for the 'CPAP pillow' - I don't have much to say about this. It doesn't seem to make much of a difference and it's not that easy to get the mask to lie in the cutaways. I'm going to continue with it for the time being. Strangely, its firmness doesn't seem to matter at all.

New CPAP machine

One is supposed to purchase a new CPAP machine every five years - at least, in Israel we are entitled to receive participation from our health insurance once every five years. The first machine that I had (not the one on the left) had internal problems after three years so I had an incentive to replace it when the five years were up.

At the time, one needed to go to an ENT doctor who would give a referral to the sleep doctor who would then write a 'prescription'. As my current machine (the one on the left) has not given me any problems, I decided not to bother with the trouble of seeing these doctors for what should be an automatic process.

On the other hand, one should replace the mask every year; I have been somewhat lax about this, even though one receives full participation and so the mask doesn't cost anything. I phoned the CPAP company the other day and ordered a mask; I was persuaded to buy a new machine at the same time. It seems that the requirement for the two doctors has been dropped: one can buy a new machine and request participation once every five years.

So I ordered a new machine a few days ago which was delivered on Friday. This is a very smart looking device although it took me a while to find where the SD card resides (the data from the machine are written to this card). Along with the machine, I also received a new mask and even a 'CPAP pillow'. 

The mask 'fell apart' when I awoke during the first night for a toilet visit, so I had to reconnect the part which separated while still half asleep and almost in darkness. This stunt did not repeat itself last night, although there are a few parts which I want to glue together on the mask. There is a part called the pressure modifier that disappeared from my previous mask and caused a terrific leak. I had to put sellotape over the holes which ameliorated the problem somewhat but not totally. Gluing this piece of plastic will prevent coming loose on this mask.

As a result, it might well be that I had much fewer apneic events over the last two nights; I can't tell as the machine records its data in a different format to its predecessor. Stored on the disk are several files, one of which being a file in JSON format (if that means anything to you, dear reader) that I can read, but I don't know yet what the various numbers mean. One can upload this file to a website that presents the data but only in a graphical manner: I want numbers!

I have written to the company that manufactures the machine in the hope that they will reveal the format of the file so that I can write my own parser and so input the numbers directly into my CPAP/walking program. At least, I hope that they will show how I can see numerical data in their web app and not just graphs; I'm a 'tables' man and not a 'graphical' one.

And as for the CPAP pillow: this link gives one an idea what I'm writing about. It's a sculptured pillow with two cut-aways (as we would call them on a guitar) that enable one to lie on one's side and have the mask fit into the cutaway. It's true that I had back pains when I first started with CPAP because of the difficult of finding a suitable sleeping position with the mask, but that was a long time ago. Last night, the pillow was harder than I'm used to, but I fell asleep and didn't even wake up in the middle of the night, so at least the pillow didn't cause problems. On the other hand. I did swim and walk a lot yesterday so I was also fairly - and honestly - tired.

The new machine's shape (think of a giant wedge of Toblerone) caused me to wonder how I could put it in a small suitcase, should we ever go on holiday again. The answer came to me this morning: leave the machine at home and take the old machine on holiday. Not only that, I can store the old machine in the small and smart suitcase which I always take with me as personal baggage: I put the CPAP machine, laptop computer, cables and medicines in this suitcase, so I have the most important things with me at all time, even if our luggage gets lost.

Masked songs

There was a full moon a few days ago, which means that the Hebrew date was the 14th of Av. For reasons unknown to me, this date marks 'the festival of love', in stark contrast the fast day of 9th Av, which commemorates the destruction of the second temple.

The cultural committee on the kibbutz decided to celebrate the locked down love festival in an unusual way: members were invited to sing a love song and act it out for the camera whilst being masked and generally incognito. The public were required to identify the singers, and the least identified singer would receive a prize. 

Even though the anonymity should mean that there would be no voting blocs (see my blog about the last kibbutz song festival), this opportunity didn't tempt me. To my surprise and secret delight, my wife, though, was interested and she had a song in mind. So I spent a great deal of time creating a suitable arrangement for this song; the key was fine, but I had to slow it down to only 70 bpm, which is the slowest tempo I have ever used. After several revisions, I developed an arrangement which met with the artist's approval.

Recording the vocals was quite easy: my wife is much better at pitching than me (i.e. singing in tune), but her timing is not particularly good. I had to count her in a few times, and even so, I can hear parts of the tune which are out of time. After mixing the song, the final track was about 3 minutes 15 seconds long. We were then told that the song had to be a maximum of 2 minutes 40 seconds, so I had to mutilate the song in order to meet this requirement. Chopping out the introduction, link in the middle and coda did the trick.

The singers were filmed last Saturday, but in order to achieve surprise, each singer was called at a different time, so none of them knew who the others were. Yesterday evening, the completed nine clips were shown via YouTube, and members had to guess who they were. This was surprisingly hard: we managed to identify three out of the other eight singers but had not a clue who the others were.

The singer who was least identified was still recognised by 32 people, which means that my wife was recognised by more than this number. This surprised us somewhat as we imagined that few people had heard her sing in recent years.

What was - IMHO - the worst song naturally won the prize for the best song. There is no accounting for taste.

The video can be seen here.

Musicians that I have heard of who share my birthday

1926 - Tony Bennett

1939 - Jimmy Nicol (temporary Beatles drummer in 1964)

1946 - John York (The Byrds)

1953 - Ian Bairnson (Pilot; plays the guitar solo on Kate Bush's "Wuthering Heights")

1963 - James Hetfield (Metallica)

There is also a musician called Kirk Brandon who was born on the same day and year as me, but I have no idea who he is.