T1g3r7th

1. Make some activity on the server. At least, to enable basic economy/trading. Have some NPC-buffers in cities [to have decent buff, at least]. Maybe, also to add some 'friend-bots', - real players will be able to create farm parties - for raids, events. And generally, if you want good server, - at least you should like it yourself, as regular player should. 2. To have Devs with skills and understanding the core things of L2 in team. Maybe, Players with experience, who know well their classes and skills. To check how this work in parties. Server w/o working sources - is not option, but waste of time. 3. The main problem of L2 - is community. New players simply don't know about the game. Today, they are downloading and installing other games. And we cannot advertise this product - simply because we not own the trademarks. So, one day, it will be much more practical - to run a new game, having best features of L2 inside, but independent and based on new graphical engine, than trying to reverse-engineer ancient stuff in a free time.

Generally I don't like when someone named himself as 'master'. There is so much hype on this topic, and with very little practical output. What I can notice about Toma: 1.When I had worked mostly on hi-level part of c6 and newer, implementing sieges on c6, functionality of clan halls, quests and so on - knowledge of key timeouts was important to build reliable server. As from 'master', I expected to hear from him some real numbers - for network connections, server timeouts, database requests, etc. With respect to C++ implementation. Not just outlandish numbers. Every person who cheated with server, will say that 10 ms - is not RTT of the tcp packet, but minimal tick time of the server. 2.Also, he failed to explain the principal difference, between tcp/udp implementations. Same about understanding of testing on single machine vs testing on real networks. Anyone who hold his own l2 server for more than year, surely know the difference. 3.About published version of sources.. A).That archives just lack - basic data stats, content of db tables - to operate the C++ build normally. Without that data, that dis-asmed piece of old garbage have no practical usage. From that point, more valuable to have java-compiled build of the server, with complete data pack - to be able to operate easily, from the single Linux machine. B).On quality of the code: Toma still can use any of old C++ standard, even which is 20 years old, still use malloc/free in cpp code and other bad practices. This will not help you to make good server, but rather turn away from making it. My advice: look in other places, ask people who have real knowledge.

I will not recommend to work with Toma, because he is rubbish and arrogant person. When I asked about RTT of TCP packet from client to server in real situation, got strange answer that it is 10 ms [which is total bs]. Same with 'understanding' of the software development process in general. There is no problems with sources, main question - is knowledge, how the things work, ways to improve. Its not here. Don't waste your time.

I have C++ code using sockets, which runs well both on Linux/Windows. Give me a reason to replace it by a legacy winsock, which will run well just on one platform? Same with the build system: I can use Makefiles in any IDE, including MSVC, but cannot run *.sln on Eclipse or Emacs on Linux. Makefiles allows every dev in team to use his favorite IDE tool, to write the code faster. You can re-compile project, install and run in one click. Even from remote PC.

I want to ask Toma, do you have interest in writing some skeleton of MMO RPG, and then attaching real pieces of L2 to it? Main idea here, to save the time of development for more creative tasks - like statistic simulation of classes, in PvP, PvE. I had worked on dozen of Java-built servers of L2, beginning from C6 and to CT 2.4. And even there, it was needed restart of all the server, after changing the sources. On debugging quests (like 4 Sepulchers), it was a long way to do it. And L2 has tons of quests. Castle sieges are very interesting to simulate, Fort sieges are interesting. I want to use total Linux-based, fully automated kernel of server, which will be easy to extend/maintain. If something special will be needed to debug, I suppose to visualize data output to external OpenGL window (which can be opened from the client's PC, just as separate program).

Sources look... kind a strange. Everything is built on winsocks and other legacy stuff. Modern C++ has threads and file system support "from the box". Also, build system... is based on M$VC IDE. Better to use Makefiles. In this case it will be possible to re-compile and run server, on same working machine, without breaking ABI. At least, good to use modern C++ for the testing env. This will save the time - especially on upgrade to C4 .. C6.