So, I suppose it is time to answer to every body...
The hosting is not the problem. Hosting is not expansive.
One more time the problem is the time consuming work to keep up-to-date the data base.
You must have in mind that, most of the time, trainzkuidindex provide links to web sites page, not direct download. Why ? Because provide direct download will violate the license of most of the assets found on freeware fan sites. In addition this help Trainz fan sites to be better known and have more visitors. Trainzkuidindex provide direct download link only if the trainz asset cannot be found somewhere else.
The downside is that the links provided by trainzkuidindex must be checked very often, because web sites change very often. Every day something change somewhere in one or several trainz fan site, even if the human user of those web site don't see any visible change. The only way to have good quality links on trainzkuidindex is to crawl the full list of trainz fan site the most often as possible (i.e every month).
You must have in mind too the size of the data collected by trainkuidindex : the current size of the collect data base is 1,7 gigabytes with more than 7 millions of lines in 22 tables. And that is only for 691 trainz fan sites! That is only for the data base, not for the files collected, witch size is about 600 gigabytes.
Some details for the data base :
- 701,539 unique kuid are referenced (found in fan sites, on DLS, as Built-in or as dependencies).
- 1,241,531 copies of those kuid that are been found in fan sites, on DLS or as Built-in.
- 3,010,108 dependencies (links between an asset copy and his dependencies).
- 199,485 links between web page and files were those assets copies can be found.
- 149,177 trainz files containers (cdp, cdpa, cmpa...).
- 703,077 links between kuid and web page where those kuid are quoted in the text (was for future use in trainzkuidindex web site)
- etc.
Checking this amount of data only by hand is definitively impossible. That is why I have developed a robot to do the main part of the job.
"So you use a robot ! Why it is so much time consuming for you ?"
Because unfortunately the robot cannot do 100% of the job, for some technical reasons too much complicate to explain here. Actually Internet is like a jungle (the rules/languages are complex and very, very permissive) and the robot is like a blind person whose job is to explore that jungle (Try to explore the jungle with a blindfold !). The robot needs an operator to help him from time to time and to check the results after it has finished his job. Crawling all the trainz fan sites currently takes between 2 or 3 weeks, mainly because of the time consuming manual part of the job (but maybe several years without robot !).
Share this work between several volunteers around the world would be very very complicate. Running the robot requires solid skills in information technology to be able to detect any problem while running and fix them. And modify the robot to be used in this way would be a disproportionate work for our small community.
The only way, if you really want that trainzkuidindex stay alive, is to accept to pay a fee to compensate the working time to keep everything working and up to date. I have suggested 15US$ per year and some people here think is reasonable. If only 75% of all people that use trainzkuidindex last year will accept to pay that fee, it will be enough to keep the site running.
This will work if every body accept the rule and don't try to cheat. In other words, if you know somebody who cannot pay 15US$ per year, don't share your user name and password, pay for him.
Pierre