Re: AI-GEOSTATS: New geostatistical open source software
I agree with Michalel, Monica and Isobel: R is a great choice! There are a lot of packges for spatial statistics and interpolation, but there is lack of more advanced techniques, like DK, multigaussian Kr, ... so, Adrian, your is a very good project!! Finally, there are R packages to build GUI (gtk, qt, java...). Kind regards, stefano Michael Grant ha scritto: SORRY PREVIOUS MATERIAL SENT BEFORE COMPLETED!!! Just a couple of comments. 1.) R has been my main programming language for a number of years now. I am involved with a mix of enivronmental tasks--statsitcs, geostatsitcs, site characterization, fate and transport, risk assessment. It is powerful, accessible and one can be productive very quickly. So much high-level functionality is already built into it, it is multi-platform and has production quality graphics. Oh yes, database connectivity is supurb. I have not grown comfortable the the geostats packages--certainly I have used parts of several of them. Sometime back noticed different variograms from one of the packages--sorry can't remember--and GEOEAS, It was nothing major, just one used the mid-point distance in the lags and the other used the average lag difference. Looking--quickly I admit--at the R-package's documentation it was clear that no discussion was there. That left me wondering how many other 'little decisions' are wired into the package(s). Still I use them, but with caution. A standard R list response applies -- look at the code ;o) The point here is that --- Isobel Clark [EMAIL PROTECTED] wrote: Have you looked at R? It is free and designed for statistical analysis. Isobel Adrián Martínez Vargas [EMAIL PROTECTED] wrote: Hello List I'm interesting to open a project for build a geostatistical open source software, with this criteria: a) extreme simple code b) Math is most important, graphic is secondary. c) Modular, as GSLIB, to make easy changes. d) Star with basic (variograms and kriging) e) End with advanced (plurigaussian, DK, or yours on methods) f)This item is for your advice... The question is what do you prefer for programming Matlab: is really easy, but it is not free... (I hope the code is free, the you can compile executable or c code in an institution with matlab license...) C, C++, know for a lot of peoples, bunt not as simple as programmers says. The advantage is that exist a lot of free compilers and toolkits as QT, glade, Visual studio C++, etc. there is also a lot of preprogramming algorithms... Python, it is really easy, and simple, it is possible to do also GUI with QT python, is platform free and interpreted language (you roon directly de source code, the system: windows, linux, etc. doesn't matter ) Other really easy programming languages can be used, as visual basic, but it is only for windows... Fortran is obsolete. TCL or Visual TCL, It is interpreted language as Python, but too slow, it is really useful to build GUIs. My Idea is make a GUI with visual TCL and make separate executables, as in GSLIB, with matlab (compiled), or use C++ for math... Finally, I was trying to play with SGEMS, but it is not as simple as it look... What is your advice? Dr. Adrian Martínez Vargas Revista Minería y Geología (Editor Principal) ISMM, Las Coloradas, s/n Moa, Holguín, Cuba CP. 83329 http://www.ismm.edu.cu/revistamg/index.htm Looking for last minute shopping deals? Find them fast with Yahoo! Search. http://tools.search.yahoo.com/newsearch/category.php?category=shopping + + To post a message to the list, send it to ai-geostats@jrc.it + To unsubscribe, send email to majordomo@ jrc.it with no subject and unsubscribe ai-geostats in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ -- Stefano Pegoretti, PhD student Università degli Studi di Trento (Italy) Dipartimento di Fisica (Physics dept.) email: [EMAIL PROTECTED]
Re: AI-GEOSTATS: New geostatistical open source software
AGAIN MY email was sent prematurely---fat fingers hitting some key somewhere. Now to finish quickly and carefully... The point is that the use of R is exceedingly attractive, but the R packages have wired a number of decisions with which you must become familar IN DETAIL if it is a basis for your package. This need for familiarity extends to contouring packages too--a separate set of issues than estimation/gridding, etc. Also note that the R package/contributor process is integal to the language's quality and success; and this process imposes significant rigor on the developer. Be sure to take a look at the material on developing packages before you make your final language selection. You will want to compare the interfaces for windows and the nix's versions--what is common between the different platforms and what is not--and look over some of the GUI-related packages (typically tk/TCL). Checkout Bioconductor and or yes there is a geospatial group within the R community (Roger Bivand in Norway?) 2.) FORTRAN is not obsolete. FORTRAN is perhaps not appropriate for your purposes at this time. The overwhelming majority of numerical groundwater codes are written in FORTRAN. The overwhelming majority of quantum chemistry codes are written in the same, many radiological codes are in the language, etc. ... On and on, much of the heavy numerical work in the physical sciences is done with FORTRAN. As a language it has been modernized, parallelized, optimized, modularized, etc. There is still a lot of F77 and earlier code around and in use but that might fade now that g77 (the GNU Fortran 77) is no longer maintained and old FORTRAN dragons (self) ride into the sunset. MUCH OF R is written in FORTRAN. There is a such a large body of everyday serious numerical crunching libraries...BLAS, LAPACK, etc. Of course if you get the GSLIB set, there is an F77 coding version and an F90 version. Is FORTRAN a good general purpose language? Well that depends on the nature of the 'general' purposes of the user :o) [OK my answer is no, but I have some acquaintances where it used for everything large and small outside office apps. Old school has been just fine, thank you.] For scientists dealing with complex numerical problems in many disciplines I would venture it is still an essential skill. The work where it is used is definitely not mainstream but is not about to fold. (Not disimiliar to another 'old' language--LISP.) Tools of the trade and all of that Such debate is another thread! I apologize to everyone for the three emails...needed to complete this ramble. Best wishes for the New Year. Good luck with the project. Michael Grant --- Michael Grant [EMAIL PROTECTED] wrote: SORRY PREVIOUS MATERIAL SENT BEFORE COMPLETED!!! Just a couple of comments. 1.) R has been my main programming language for a number of years now. I am involved with a mix of enivronmental tasks--statsitcs, geostatsitcs, site characterization, fate and transport, risk assessment. It is powerful, accessible and one can be productive very quickly. So much high-level functionality is already built into it, it is multi-platform and has production quality graphics. Oh yes, database connectivity is supurb. I have not grown comfortable the the geostats packages--certainly I have used parts of several of them. Sometime back noticed different variograms from one of the packages--sorry can't remember--and GEOEAS, It was nothing major, just one used the mid-point distance in the lags and the other used the average lag difference. Looking--quickly I admit--at the R-package's documentation it was clear that no discussion was there. That left me wondering how many other 'little decisions' are wired into the package(s). Still I use them, but with caution. A standard R list response applies -- look at the code ;o) The point here is that --- Isobel Clark [EMAIL PROTECTED] wrote: Have you looked at R? It is free and designed for statistical analysis. Isobel Adrián Martínez Vargas [EMAIL PROTECTED] wrote: Hello List Im interesting to open a project for build a geostatistical open source software, with this criteria: a) extreme simple code b) Math is most important, graphic is secondary. c) Modular, as GSLIB, to make easy changes. d) Star with basic (variograms and kriging) e) End with advanced (plurigaussian, DK, or yours on methods) f)This item is for your advice The question is what do you prefer for programming Matlab: is really easy, but it is not free (I hope the code is free, the you can compile executable or c code in an institution with matlab license ) C, C++, know for a lot of peoples, bunt not as simple as programmers says. The advantage is that exist a lot of free compilers and toolkits as QT, glade, Visual studio C++, etc. there is also a lot of
Re: AI-GEOSTATS: New geostatistical open source software
Hi Michael Nice to see someone comfortable with rambling. I think we should have more of that in the list! Being an old warhorse and too far gone to change, I still use Fortran. My excuse is always if it's good enough for NASA.. Visual Basic is pretty good too but Fortran is still the faster 'computational' language and a very easy interface. I guess we'll just have to phase out as the compilers disappear ;-) Isobel http://www.kriging.com
