The ECIM panel session

Editor Neil McNaughton resumes and embellished his contribution to the ECIM panel session. What were the great data management fails of the past 25 years? The great successes? And why are we still asking the same questions anyway?

In my opinion, ECIM continues to be the best upstream data management conference*, a position it has consolidated with the demise of the other ‘best’ show, PNEC. It was gratifying that the organizers gave me some decent exposure with an appearance on the panel session and some amusing, if rather embarrassing probingduring the gala reception, by ECIM court jester Graeme Blakey, of my failing memory .

One thing about being on a panel (at least for me) is that one is so focused on ones own discourse, trying not to say anything too daft, that it is hard to listen to what the other panelists are saying. To remedy this I invited the other panelists to provide a short summary of their key points. So far nothing, but the invitation is still open. So that leaves me with the opportunity of summarizing and, why not, embellishing my own contribution. Which, in other words, is pretty much like writing an editorial. So here goes.

One thing I have noted from the many previous panels I have listed to is that panelists often do not seem to be addressing the questions from the master of ceremonies. This might be frustrating for the MC and those in the audience who are expecting some serious answers to good questions. Well I’m not sure which question I was trying to answer, but I first found myself extemporizing on the topic of standards, divulging a key argument form my as-yet-to-be-written book which may or may not have a title along the lines of ‘The IT that Goes Wrong’.

So standards! I have been attending conferences, user group meetings and what have you for many years and have observed two things. One: People entering the profession confronted by a problem of data exchange or interoperability all come to the same conclusion that what is needed is a (new) standard. Two: Following possibly many years of effort, the same folks discover that the new standard is not working and move on to other things, to be replaced with new blood, folks with renewed enthusiasm and the process starts over.

In years of fiddling with my web development, working across platforms (PC, Linux, Mac) I made a major discovery in the standards arena. As you know, the king of all standards is Ascii, the American Standard Code for Information Interchange. This sets out how text and numbers are represented by bits and bytes. But when you are working cross platform with Ascii you will very quickly come unstuck with the manner in which end of lines are handled across the different platforms. Given that Microsoft and Apple are competitors (as is Unix/Linux) it is easy to see that the intent is to hamper interoperability. After all, the last thing that Microsoft wants (and vice versa) is to make it as easy as possible for its user base to up-sticks and transfer all its workload to a Mac environment. The EOL gotcha is real, but it is also a metaphor. Similar gotchas exist between cloud platforms, printer cartridges and just about everywhere in IT. The problem is for those advocating standards is that there are far too many people pulling in the opposite direction, and these are the people with the clout!

All this was by way of softening up the audience for my bombshell. This was in answer to the question, ‘what has been the greatest data management disaster?’ I found myself saying ‘the semantic web’! Which should have caused a riot in Norway, home to ISO 15926 and a host of academics beavering away to leverage Tim Berners-Lee’s (a.k.a ‘God’) resource description format to make a better world of interoperable linked data. Well that hasn’t happened. In fact it hasn’t happened so much that my bombshell went down like a lead balloon. I doubt that many in the audience were aware of the years of semantic effort**. It is true that semweb has been more popular (but essentially unsuccessful) in the engineering community than in the upstream.

Sitting next to a SLB rep, I thought I would give a smart-ass answer to the question, ‘what has been the greatest success in data management?’ Schlumberger’s Finder I said. Which was ‘smart’ in that Finder has long disappeared from the SLB catalogue, replaced with a host of expensive apps that are now to be replaced with OSDU (maybe). It was also smart because for a while, Finder really was something else. Selling like hot cakes in the late 20th Century. It was eventually decommissioned when the underlying software stack reached end of life (more IT that goes wrong!).

I also found myself giving another smart-ass answer to the ‘greatest success in data management?’ question. I suggested the greatest success was GIS. Smart, because for all the efforts of the bottom-up data managers to manage E&P’s plethoric data types and re-combine them into a prospect or play, it seems like the top-down GIS approach often provides an easier way of bringing it all together. Perhaps I should have said, that GIS is better at stealing the scene from the data managers who still need to do the grunt work getting stuff in place.

The final two questions were ‘What is the biggest unsolved problem in data management?’ and ‘What are you most hopeful about for the future?’ Tricky ones these. On the one hand, the biggest ‘unsolved problem’ and their hoped-for solutions have remained remarkably static over 20 plus year of data management conferences. I think that on top of these, a new problem is the siloization of data management and the difficulty of communicating between the data managers, IT, data science and ‘the business’. So my hope was to do this better. But since IT is constantly mutating, rediscovering the wheel and inventing new names for stuff, this is not getting any easer. On the subject of naming stuff, another divulgation from my future oeuvre. The great French mathematician Henri Poincaré defined mathematics as the ‘art of calling different things by the same name’. I suggest that IT can be considered the ‘art’ of calling the same thing by lots of different names.

* Of course the other (the real?) reason I like ECIM so much is the opportunity to run in the beautiful Djupadalen park. You can see my 2022 ECIM run on Strava.

** The semantic dream lives on though. In this very issue we report on how Australian academics are still flogging the semantic dead horse in an MRO context.

This article originally appeared in Oil IT Journal 2022 Issue # 6.

For more information or to comment on this topic email here.