No apologies required or expected.

1. No hurry. I have recently found myself under an enormous pile of work.

2. It would be foolish to start in any other way. Get your bearings first. Sorry about the broken link. I noticed it earlier when considering a reply to Steve McIntyre and his admonition to use R.

3. Glad to see we are in agreement here. I have, when possible, written the core as a library, dll, or driver, and put a gui or scripting interface on the front end. I did not follow MS over to .NET and C#. Not because of an aversion to it the opportunity just has not presented itself. I am currently writing web services in Java.

c

I apologize for the slow response. I’ve been busy, not “waging war”, but engaging in little skirmishes.

It’s late but I will try to quickly answer your questions:

Did I know what I was getting into?
Nope. I wrote a quick parser for GHCN data, then very quickly evolved it into a geographically weighted averaging program. It fills a useful niche and it’s gotten more attention than I expected. The ability for anyone to generate results is powerful.

1. Functional Requirements / Use Cases:
I haven’t yet stepped back to think about the big picture. I have some use cases in mind. I’ll write you offline with more info.

2. Steve Mosher:
The link was to the article (not a comment). Did you mean this one about OpenTemp becoming the standard?
http://www.climateaudit.org/?p=2200.#comment-149524

I didn’t have such lofty goals when I started. With a flexible design and great documentation it could become the standard.

Right now there’s nothing magical about it and to be honest it’s structure is a little embarassing (tested and well commented but quick and dirty).

3. Architecture/Design Guidance:
Although I’m a Windows/C# guy, I like the Unix mantra of sharp tools for a single job. I envision building it as many sharp tools (for parsing station data, calculating cell temps, geographic averaging, graphing, mapping, etc) that are linked together with higher-level programs or scripts. Some of the tools should be wrappers around existing software.

I will stop here before I ramble on much further. Let me know your thoughts.

Thanks for the clarification. I did not know how to interpret “forward-engineering” in the context of the Hansen reverse-engineering thread. I could not decide if were you re-engineering or that the analysis showed a need for a new set of requirements, etc.

Motivation:
On one hand I am still happy to sit on the side-lines. On the other hand I would like to provide researchers (like MikeM) a tool that can help sort out the extent of the warming, the anthropogenic component, and confidence of each measurement. Since you got there first I have been content to sit back and watch– did you have any idea of what you were getting into? At present I am willing to supply arms (for all sides) but not willing to engage in war . You seem willing and able to do both.

In reply to your query:
The project is interesting– wish lists are just work to be done I will download the latest version of OpenTemp and take a more serious look at what it is doing and draw up a few documents to further my understanding.

Questions:

1- I realize OpenTemp started as a weekend project but did you have a set of functional requirements or use cases in mind?

2- http://www.climateaudit.org/?p=2200. Not that Steve Mosher speaks for you but would it be correct to assume that this is the general direction OpenTemp is headed? Can you elaborate?

3- If relevant, what use cases should be considered to accomplish #2? I see your wish list but I am looking for architecture / design guidance. It seems pointless to re-invent statistical computation, spatial, and graphing engines but it seems OpenTemp has to be both input and output to each.

It’s a common misconception that OpenTemp is an attempt to port and modernize GISTEMP. Actually, OpenTemp is an independent (and simple) program to calculate temperature trends.

Currently it is artificially restricted to the USA lower-48 with hard-coded analysis boundaries. This was intentional to prevent its casual use in other areas of the world before its limitations on the well-studied USA48 data are understood.

I would like to start moving out to the rest of the world before the end of the year. I also have a long wish list for OpenTemp in the following areas:
- improved, but still very simple, architecture
- improved input formats
- more robust geographic averaging
- pure trend analysis (no absolute temperatures)
- chart and map output
- proper website (probably Drupal)

If you would like to help, that’s great. Do any of the above areas catch your interest?

I have many other questions but this topic will get me to my first decision point– I am also a software developer and looking for something to occupy the free cycles I seem to have during the winter months.

ps– I am not a stats guy. In fact, I absolutely hated stats in school

You can absolutely use the program for that purpose. It will soon be properly open-sourced so that anyone can use it for anything.

I sent you an email with more info.

BTW, the newest version is now available here:

http://www.opentemp.org/_release/v1rc1.zip