Have you read through Applied Energistics code? Have you read through Factorizations Router code? Have you tested the server load on a large scale? Didn't think so. No I have not. But I have meddled in my fair share of mods and have also made private mods for servers, and maintained a well known public one for a little bit. I know full well what code is needed. Whether it is 1 block or 50 blocks interfacing with 50 machines the code needed to run the interface for 50 machines is pretty much the same, the only difference is the single block has to search and check all attached inventories, and also constantly check to see if there has been any change to the connected machines, like adding or removing one. In fact having only 1 block searching and meddling with 50 machines could, in theory, be quite detrimental to performance, especially if used on large numbers of machines. The whole performance debate came up with a simple statement. The server is already under stress. The performance is already taking a massive hit with the mods we have, why not wait and see if performance can be improved before adding one that can possibly hurt the performance more. All of this is a moot point anyway, in the end SK said he was considering it and ultimately no matter how much we bash heads here, it is up to him.
I'm still on the fence about Factorization. I would like to assume that it will be a positive addition to the server, but without testing/evidence we can't know for certain. As a point of curiosity. Isn't it more efficient for a single block to interface with 50 blocks than to have 50 blocks each interfacing with 50 more blocks? 51 vs 100 blocks that have to checked? Or, Is it the same because the single block gets checked 50 times anyway?
It really all depends on how each of the blocks does the checking; one block checking 50 blocks in an inefficient way would be much worse than 50 blocks checking 50 blocks in an efficient way. I don't really know which is better in this case as I haven't looked into it at all, but, unless the person really screwed up, they should be very similar in terms of performance.
I know this Ravens but there are other things more important then Factorization that needs to be dealt with. I wasn't saying he said No to the whole thing, rather No to it being installed at this time.
Yes but the server as a whole would rather have AE then factorization. AE is simpler to use and stores everything into a machine and a disk for easier use. Yes Factorization is nice but AE is nicer. In my honest opinion.
Another thing to consider is that export buses need to constantly check for items to export, where a router idles if it's inventory is empty. An export buses only "idles" if its destination inventory is full. I decompiled, and checked to confirm this. AE looks very well written and efficient, but the amount of work involved with each export is considerable, in comparison to anything a router has to do. Unless there is a major bottleneck (profiling would be needed), routers would have to be much faster. The only heavy lifting a router needs to do is update it's machine grid, and it does that rarely.
