[Edserplo_users] Apologies for delay in fixing bugs

[Steve Loch]

Apologies are due to various users for not fixing various minor but important bugs (like an incorrect vertical bin axis for 2D data). This has to do with a major restructuring of the code currently being undertaken

a) to make it more intelligible and maintainable 

and 

b) to allow Edserplo's dump and settings files to be written in NetCDF. 



On (a) we had a professional contractor in at one point last year and he could make no headway at all. This is no doubt a reflection on his talents but it is also an indication that the code is over-complicated for what it does. There has been a failure to properly decompose the functionality and also a willingness to duplicate code leading to poor maintenance characteristics (and unresolved bugs) in certain key areas.

To understand the significance of (b)  you should know that dump and settings files are Java-dependent. This means that changes to the program can render past D & S files illegible. Such changes can be very minor - like adding another function. In other words the situation  puts a brake on development and is inherently unsatisfactory. Using NetCDF means theat the contents of these files are immediately accessible and largely independent of coding changes in Edserplo or in Java releases (or indeed NetCDF API releases) . Files are likely to be readable for periods to be measured in years or even decades rather than months.

There is the question of why it was not done with NetCDF originally: primcipally because doing it the Java way (referred to as serialisation) was easier and quicker to implement and importantly guaranteed fast IO. It also took Ravi a long time to come to grips with the NetCDF Java API.

These changes should be virtually transparent to users of course. I'm working on the idea of making old settings files legible for a period but any update will overwrite them with a NetCDF version. 

Finally I should say that (b) is a necessary condition for Edserplo to take over from Edteva.

------------------------

If I cannot to produce a new production version with its hundreds of changes in the near future I will have to fix the minor bugs using the production version of the code. There is one significant hold-up at the moment which is why I prefer to concentrate on it in the hope that it will be resolved soon.  

Steve