What it would take to build VEF.....

[TiG]

First off - i don't profess to know it all about this, and i will be asking to people to add points but i'd like to keep this as clean of comments as possible so please respond with ideas thoughts to this thread...

http://forums.hexus.net/showthread.php?t=20024

Right you need a front end to a VEF replacement.

1) My thoughts on this based on my experiences are to remove the choice of mission and auto generate the lot.

Reasons for this is based on a few things - people then can't get their choice of plane, remove the BF110C issue with OKL and try and make the plane use based on more historical figures.

Plus technically it makes it easier to manage.

2) so you have a web site - do you need to have logins?? surely this is governed by Hyperlobby log in?, or is it?, i've not seen any reason to have a login on the VEF site.

Track based on Pilots reported in missions, Those that don't fly within 1 month of last reported mission will just be dropped from the database.

3) So mission generation would have to be similar to VEF, you want to host then you have to enter a part of the sight and it will down load you a mission. Some form of user Identification would be used here as reports of bad hosting cheating etc etc would result in a bad from being able to host.

I want to host a mission i get auto generated file. Hosts must be responsible enough to only request missions when intending to fly them.

4) Mission generation believe it or not is probably pretty easy.
I could write the text file myself using a database generated system, or as the case seems use DGen or even replace it with something liek this...

http://www.lowengrin.com/content.php?article.12

5) mission reports happen same way as they do now, upload of co-op.txt - which then parses the results and inputs onto the database to allow querying on.

Technicals to follow of how to do this?.

[TiG]

Right how to set up the structure for this - firstly being a proper developer out in the wide world of business, I for my Sins am very competent with MS SQL 2k. It would be my DB of choice apart from the cost. But then as i'm not actually planning to implement it matters little.

So i'd build this thing using 2 Databases. Why 2 i hear you ask. Well for a start, where you've got a system updating lots and lots of records with missions being completed the last thing you want to do is everything thrashing the database for how they did on their score.

So it writes to one, reads from the other, replication is easy within SQL2k and suits this well.

So thats the reporting side done.

The Web site, obviously for a business you would buy 2 web servers and run them in load share setup. Same goes for this, both web servers only connect to the read driven database.

Mission Generation is done at 00:00 midnight with a chron job that generates say 200 Missions, (adapt as necessary) - web site logs which mission is next, and logs which host requested the mission.

People play mission, get shot down die etc and report mission.

The mission reporter for me would be an asp upload page which takes the co-op file - checks for the name of the file it expects for this host and parses the combat strings.

It logs what you killed and stores it in the write only database. Which then replicates across. It also means that generation and reporting should not stop if people are thrashing the web database to get their stats

Web database is setup purely into a easy query format in 2nd Normal form to allow efficent gains.

I'll restructure this to make it a bit neater later....

TiG

[TiG]

P.S if you got bored, lost or confused, take it up with ZAK its his fault

TiG

[Zak33]

THIS THREAD IS VITAL...as discussed...dont clutter it up people....just keep it clean and concise

Same rules for the other one...which is stickied

[Neil Washbrook]

TIG - can I suggest you take this to the VEF development forum where there is also another fairly well advanced discussion going on about replacement.

[Hat]

http://www.netwings.org/cgi-bin/dcfo...onf=DCConfID11