Thursday, May 24, 2007


A few weeks ago I received an email via the SoundClick site: someone wanted to 'be my friend'. I contacted this person and discovered that she lives in a suburb of Tel Aviv! The whole world uses the internet and I get contacted by someone who lives maybe 30 kilometres from me! Anyway, we started corresponding and one evening I idly suggested that we collaborate on something.

After she told me that she doesn't write, I suggested that I create a music track for her and that she sing it. After a bit more to-ing and fro-ing, we agreed to do a version of 'Halicha LeKesariya' ('Walking To Caesariyya') which is colloquially known as 'Eli, Eli' ('My God'). This song was written by Hanna Senesh, a young immigrant to Palestine (as it was then) who volunteered to parachute into Nazi occupied Europe in 1942 and was never heard from again.

I wasn't too sure how to approach this song, as it's very much a classic Israeli song; due to the sombre meaning which has been attached to it, I felt that a very simple arrangement was in order. My original intention was to have a very dark and engulfing pad against which would be set the crystal vocals of my collaborator. We did indeed make such a version, but I got tired of the sound and decided to remix the music, changing the instruments. I then sent the music file via a file sharing website (essential for collaborations) and was sent a version back complete with vocal.

I didn't like the sound of this version, and a quick look with a wave editor showed that there was a large dynamic range: the vocals were creating very loud peaks and the music was very quiet. After a few more emails back and forth, I requested that I be sent the vocal file so that I could make my own mix. I had to apply very heavy compression to the vocal and some eq, but this enabled me to make the music much louder and better balanced.

Both of these versions can be found at my collaborator's website.

After having spent a few hours on the remix, I decided to take the dog for a walk. On the way, I was listening to some of my instrumental pieces, and I realised that the song would sound better if I ditched the current arrangement and swapped it for one of my patent arpeggio based tracks. So when I came back from the walk, I started yet another arrangement. I kept about 50% of the original and then added two electric pianos playing arpeggios, one on each channel. The second piano is playing the same notes as the first, only delayed by one beat, thus producing an interesting effect. Another couple of hours working on this, getting everything just right, and now a third version of the tune was finished. I think that this one is the best; it's more open and sounds much better.

One part - a high pitched violin or voice, it's difficult to tell - is playing notes which are generally a fourth or a fifth apart. Thus the part starts with an A, jumps up to D, then up a fifth to a higher A and then down to E. I wasn't really thinking about it at the time, but that second A is played over a Bb chord, and thus is a major seventh; it sounds very cool. The chord sequence which I used is one slightly more adventurous than most people might use to accompany the song, and includes a very interesting sequence - F Bb Em A. The transition from Bb to Em is thrilling.

For the time being, I'm putting this third version up on my SoundClick site.

In the mean time, I've been sent another song to arrange. I've listened to this only once, and it was absolutely dreadful, featuring a guitar which sounds totally out of time. It's going to be interested creating a track for this, as first I have to learn the song and that's going to be hard. I'll just try and write down how many bars there are of each chord and start from there.

Tuesday, May 22, 2007

Metadata and information

I wrote a few days ago about reports which give statistical data. These reports provide metadata: data about data. Such reports tend to be more interesting and higher level than reports about 'ordinary' data, and in fact management should be interested only in such reports, but there is a natural tendency to sink down to the data level and start checking each specific order instead of looking at trends.

I participated in a course about knowledge management once; one of the simplest things to come out of this was the following hierarchical model. At the lowest level, one has data - eg specific orders. Above this level is information - eg what percentage of orders are being delivered on time. The highest level is knowledge - how to create conditions such that all orders are delivered on time. The transition from data to information isn't very difficult, but the transition from information to knowledge is much harder, and that is the reason why companies try to hang on to employees who have knowledge.

The course itself was about the formal storage of knowledge within an organisation: how to know what knowledge people have, how to formalise that knowledge and retain it, and even more important (my project in the course) how to know what knowledge does not reside within the organisation (that's where I first used the theta query). Unfortunately the course fell prey to problems within the college where it was taught, and so it didn't fulfill its initial goals.

Is it surprising that I work in information technology and not data processing?

Today is my best friend's birthday (you know who you are!) and tonight marks the beginning of the Shavu'ot festival.

Monday, May 21, 2007

Train journey

I spent Thursday giving training sessions to users (both new and advanced, although not at the same time) with our ERP program in Haifa. This was cunningly arranged so that I could spend Thursday evening with friends in the north of Israel and then return home the next day without causing any problems at work (our work week is Sunday - Thursday, although I always work a few hours on Friday). As fortune would have it, there is a train station near where I live and one very near where I was giving the training sessions, so it was a no-brainer travelling by train instead of car.

On Friday morning, I was taken to the train station in Nahariya (11 kilometres from the Lebanese border) and boarded a train which was headed for Ben Gurion airport. My intention was to alight at one of the stations in Tel Aviv and then change to a train from Tel Aviv to my station (Bet Shemesh). But once the journey got underway, the conductor said that the train would continue past Ben Gurion airport, stopping at Lod, Kiryat Gat and Be'er Sheva. These names probably won't mean to non-Israelis, but the point is that the train was leaving the most northern station of the rail network and continuing through to the southernmost station - a very long journey (it's like a train I once boarded in Britain which went from Liverpool to Bournemouth).

I was very surprised at this information as I wasn't aware that the network had been extended past Ben Gurion airport. The airport had only been added as a station about a year ago and I had yet to travel there as it was the end of a line. But for the past few years I have been watching in awe as a bridge had been built by the airport, arching over the motorway from Jerusalem to Tel Aviv. Step by step this has been a fascinating process to watch. Listening to the conductor's announcement, I realised that the bridge had become operational, and so decided to stay on the train until Lod, where I could switch trains.

The track from Tel Aviv to Ben Gurion is short and mainly underground, arriving at the basement of the airport. This line could be very useful for people flying, although of course it depends on the hours that the train runs. Then we continued out of the airport and over the main road - not as fast as I would have liked, but still impressive. A few minutes later we were in Lod, where I alighted.

I suggested later to my father that he spend a few hours having fun on the railway (he doesn't have much else to do): take a train from Bet Shemesh to Lod, change and wait for a train going to Ben Gurion and points north. Change at Ben Gurion and then come back. It's good for a morning's entertainment and won't cost very much.

I still haven't travelled to Jerusalem by train, something which is more enjoyment than necessity. Apparently the views are astounding (and having walked several times along that region, I can appreciate this), but the journey is slow.

Wednesday, May 16, 2007

The 'theta' join and SQL arcana

A good day at work yesterday and the day before. I managed to solve a syntax problem preventing the execution of a new report, and in doing so learnt a lesson which I had been taught before and forgot. This gave me renewed confidence regarding the report generator in our new ERP program, and so yesterday I was able to create a few new reports.

These are statistical reports, showing how long the order approval process takes, how we stand regarding supply dates (as opposed to what we promised the customer), how long it takes to repair a chair and how long it takes to complete a service call. The quality management system (which I used to run) has goals for each of these services, and it is important for us to know where we stand in relation to these goals. In case we should surpass them, it's always possible to move the goal posts.

Coming off these reports, I had an 'a-ha' moment when I saw how I could write a report which shows which parts don't have a bill of materials. This can be generalised to "give me a list of which doesn't have ", for example chairs which no one orders or unordered upholstery colours. I read somewhere once that this is called the 'theta' join: how to find values which are in one table but not another. In 'classical' SQL, this can be written as

select, customer.ip
from customer left join subscrip
on = subscrip.customer
where subscrip.ip is null
(this is an example from a program I wrote which manages the billing for internet access on my kibbutz and will find end customers who don't have billing data).

Unfortunately this syntax doesn't work as a 'naked SQL' query in the ERP program, let alone translated into the version needed for the report generator, so I had to shelve the idea originally. Yesterday I remembered seeing a query for a report which someone had written which used a subquery; I realised that I could use this as a replacement mechanism for the theta join. Translating back into SQL, I wrote
select part.partname, part.partdes
from part
where not exists
(select 'x' from part p1, partarc pa
where p1.part = pa.son
and pa.part = part.part)

What is this saying? If the subquery returns a value (that's the select 'x' part), then the where clause will be false and so there will be no value for the query. But if the subquery fails, then the where clause succeeds and so there will be values for part.partname and part.partdes.

A large step for (New) man, a small step for mankind. Or maybe it's the other way round: an hour's work from me and we can now produce data which other companies can only dream of.

Saturday, May 12, 2007

C. Arthur Milton

I learnt today in a roundabout way of the death of C. Arthur Milton:

"In his time, the footballer and cricketer Arthur Milton was as big a name as Alan Ball, the World Cup hero whose death preceded his by a few hours. Milton played on the wing for Arsenal and England and opened the innings for Gloucestershire and England. As such he was the last of the double internationals." - taken from his online obituary.

Normally, the death of such a has-been sporting star wouldn't mean much to me, but Mr Milton was special to me, as he was the father of Robert, my best friend in junior and secondary school. I wrote a little about this here.

I didn't exactly "finish" secondary school, I sort of "evaporated". In those days, there weren't any lessons in the summer term for the upper sixth form; pupils stayed at home and revised for their A-levels, coming into school only to sit the exams (or play cricket). My parents moved from Bristol to Cardiff before my exams had finished, and so I was a very rare sight around the halls in the summer of 73. As a result of our move and my spending the 1973/4 academic year in Israel, I lost touch quickly with almost all the people with whom I went to school.

A few days after coming home in July 1974, I took a train to Bristol and met up with Robert again. We spent a pleasant day talking and making music, and then went on our separate ways again. We next met up about a year later when Robert came to see me in London; he came loaded with a guitar and an amplifier which I had to shlap from West Hampstead tube station to my London residence. Again, we spent a weekend making music together and then went on our ways.

I suspect that I wasn't too hospitable towards him (a charge which had been levelled against me by someone else at around the same time regarding another event) or maybe we had just grown too far apart (I was quite different from my old schoolfriends by virtue of having spent that year in Israel, being dedicated to emigrating and being gung ho about my youth movement), but for whatever reason, we never got in contact again.

I have tried over the past few years looking for Robert over the Internet but have not been successful. Even today I tried and failed again, although I did find the address of the Milton family home in Bristol. I didn't remember the name of the street, although when I looked at the address on a map, the old neighbourhood came into focus, with my old house at the top of the map, our primary school in the middle and Robert's house at the bottom.

I'll send a condolence card tomorrow along with my email address and we'll see if anything transpires.

Wednesday, May 09, 2007

Killarney Boys Of Pleasure

I uploaded a few new instrumentals to my SoundClick page, including a traditional Irish tune bearing the very politically incorrect name of 'Killarney Boys Of Pleasure'. I was in Killarney once but I didn't meet any boys of pleasure (and probably would have run a mile had I done so). I imagine that the meaning of "boys of pleasure" has changed somewhat over the years....

Anyway, this is normally a fast paced violin tune, but the version which I know by Lunasa slowed it down "to make it sound groovy", to quote the sleeve notes to their "Merry Sisters Of Fate" disc. Their version informed and inspired my version, and to give credit where credit is due, I think that their version is better.

I saw Lunasa when that disc came out, and reviewed both their concert and the disc for the GreenMan review online magazine. I couldn't find those reviews now after a quick search on GMR's website, but my review of then Lunasa guitarist Donogh Hennessey's music book is still there. I thought that I was doing the band a favour by giving this book a positive review (and I bought it with my own money), but Donogh sent a vicious and personally insulting reply in return which totally turned me off reviewing.

When in Ireland in 2002, I bought their followup disc, "Redwood", but didn't like it. I still don't like it, not for personal reasons, but for professional ones: the dynamics which made MSOF so good are totally lacking in "Redwood" and turns Lunasa into "just another Irish band".

Anyway, I digress (and what's the point of blogs if one can't digress every now and then?). My version of 'Killarney Boys Of Pleasure' is up at SoundClick, I hope that you'll listen and give it a 'hot' vote.

Tuesday, May 08, 2007

(Welcome to the) Hotel California

I've been reading "Hotel California" by Barney Hoskyns over the past few weeks. I purchased the book because I had been led to believe that it contained a fair amount of information about Randy Newman (and because British Amazon had it on sale for a ridiculously low price) and some about Jackson Browne. It turned out that there was plenty of information about Jackson Browne but little about Randy Newman, who didn't really belong to the Los Angeles singer-songwriter circle of the late 60s/early 70s.

This book talks about a period of music which isn't well represented in my collection. Apart from the Newman albums, I have all the early Jackson Browne albums along with "Crosby, Stills and Nash", but my Joni collection starts with "Blue" and concentrates on the following years. Once I bought a 'greatest hits' collection by The Byrds, which I listened to on Saturday. Almost every track sounds the same and almost none of it sounds interesting.

This is a concise description which I could apply to most of the music described in Hoskyns' book: most of the faux cowboy material fell on deaf ears and I never could stand The Eagles. As someone rightly pointed out, almost all of the musicians which one considers to be part of this scene are not Californians (the exceptions are Newman, Browne and Crosby).

The book is a bit muddled, unsure of whether to tell a strictly chronological tale or whether to concentrate on people and/or groups. As a result, there's a bit of both which makes for haphazard, if informative, reading.

I knew nothing about the early life of David Crosby (all I did know was his sailing, his drug problems, his liver transplant and his jail sentences), so this book clued me in. To me, the standout track of "Crosby, Stills and Nash" is C's "Guinevere" and I have always wondered why there were no other tracks of a similar standard created by him. Now I know.

I have the "remixed from the original master tapes" cd version of CS&N which I bought about ten years ago. The production on this is strange to my way of thinking: whilst the guitars (especially in "Suite: Judy Blue Eyes") have careful spacial placing, the vocals (with one notable exception) are mixed together, dead in the middle, thus making it difficult to separate each singer's contribution. Strange for a harmony group - unless they wanted to provide a blend of three voices as one. The notable exception is "Helplessly hoping", which has clean separation on the vocals (and showing again that Crosby has a lovely voice); this track only has a single acoustic guitar, which may have something to do with the vocal separation.

I've never been one who looks for the perfect vocal (I have been given to understand from other sources that both CS&N and The Eagles would record multiple vocal tracks and then edit on a word by word basis!), as one can imagine from listening to years of Randy Newman, so most Californian perfect vocal songs leave me cold. Where is all the expression and emotion? Rank The Eagles against Van der Graaf Generator, and whilst it's easy to understand why one act has sold millions of records whereas the other has not, it's also clear on which side the true emotion and passion lies. And the musical ability.