 |
|
A Desktop Experiment
Management Environment
 |
|
|
| A Desktop Experiment
Management Environment
|
Horse: Object-based database server.
Turtle: Translator between Moose objects and Ascii files
Emu: Experimentation manager
Tutorial: A Sorting Experiement
- The Version 1.1 release of Horse is supported for Sun SPARC Stations and Pentium Pros running Solaris 2.6.
- Horse uses Informix Online Server 7 as its storage manager, so you have to install Informix to run the released binary code. You can also build your own Horse on other Relational database servers. For this purpose, we release source code of the interface module between Informix and Horse, with the rest of Horse being given as a library. By modifying the interface module appropriately and linking it to the library, one can obtain Horse versions for other RDBMSs.
- Our compiler is egcs-1.0.3 for C and C++ and INFORMIX-ESQL Version 7.10.UD1 for Informix ESQL programs. Although we have made an effort to ansi-compatible, there are places where there are inconsistencies.
- Install and setup the Informix Online server.
- Download the release of Horse binaries.
- Unpack the package, and follow the instruction in the horse.ps file inside the package on how to set up the environment and start Horse. Refer to the "Moose and Fox User's Manual" or to the original Moose-Fox paper "A Moose and a Fox Can Aid Scientists with Data Management Problems" for the syntax of Fox.
The Zoo Desktop Experiment Management Environment currently consists of three components: Horse, Turtle, and Emu. One can download each component seperately or all of them together in a single package. There is a document in the list of manuals for each component as well as for how to set up an experiment using the entire system.
Horse is designed to work along with Turtle, Emu, and any preexising external software to help scientists conduct experiments. The workflow of an experiment can be specified through a Moose database schema.
We have released a tutorial experiment (a simple sorting experiment) that shows how one can specify the experiment workflow in a schema and write some agents to connnect the external experimentation software to Zoo.
One can either download the binary code for the experiment or download its source code together with all necessary libraries. After unpacking the downloaded package, one should refer to the readme file on how to run the experiment. One may also want to download the documentation on the Tutorial Experiment.
Build Horse to Work on Top of Other RDBMSs
The current version of Horse has been built on top of Informix. We do intend to generate versions of Horse on top of the ODBC standard for other relational engines. For the time being, to allow usage of Horse on top of other systems, we are releasing the source code for the lowest, Informix-dependent layer of Horse (together with all necessary libraries and header files for other modules), which people may want to modify. This module has been written in C++ and the Informix Embeded SQL. One should download the above source code and refer to the readme file in the package on how to compile it.
Please contact zoo@cs.wisc.edu
if you any have comments or questions.