CommunityTSC/Seal of Approval

Aus Rechenkraft
Wechseln zu: Navigation, Suche

CommunityTSC Gütesiegel engl.png

Evaluation of CommunityTSC (May 2008) (Deutsch ja.gif    Deutsch)

Evaluation took place in 4 categories, each awarding credits ranging from 0 to 5 points, and an overall rating. We had employed an evaluation board of 3 people, the (rounded) average was used.


Category 1: Reliability

There are clients for Windows, Linux, MacOS X and Solaris. The client has a very nice GUI, but there is also a CLI client available for Solaris and Linux. But both requires Java. However, it is a very serious shortcoming that there is still no checkpointing after so many updates, and because of this the whole effort of an unfinished work unit will be lost at a restart. The servers are good accessible and there is enough work.

CommunityTSC scores 3 points in this category.

Category 2: Communication Policy

CommunityTSC operates a fairly large forum with a lot of news stuff, while there isn't much informations on the actual homepage. Answers of the project operators come very late and mostly the answer isn't much informative. A correct description of how exactly the project works and what the client actually calculates is apparently absent.

CommunityTSC scores 2 points in this category.

Category 3: Instructions

Instructions for the different clients are available in the download section. The how-to for Windows is very well explained, while the instructions for Linux or Solaris are sparsely. Because of this, beginners could have problems running the program.

CommunityTSC scores 3 points in this category.

Category 4: Benefit for the general public

CommunityTSC in itself is an interesting project with an important goal and great benefit to the general public. The project is crunching now for many years to provide the best candidate. But these haven't been tested yet under laboratory conditions because the project operators hope to find still better candidates. This doesn't make any sense, since they will always find better candidates. Due to this fact they wasted already thousands of CPU hours, because after months they found out that a flaw in the Work Units exists. As soon as the operators begin to test good candidates, it would be reasonable to go on crunching for CommunityTSC.

CommunityTSC scores 1 points in this category.

Overall Rating

Parts of the project are very well done. The client shows the approach of a very good project. But unfortunately this is all useless if the candidates won't be examined under laboratory conditions.

CommunityTSC therefore scores 2 points over all categories.

Eigene Werkzeuge