Shareaza

I welcome. I am Polish, poorly I am good at English therefore I am placing the computer translation from Polish to English of the text with my proposal to improve Shareaza.
I propose that it is possible to send priorities host which Shaeraza is merging according to tongues in which we are making available to or we are seeking files. Moreover I propose that searching is also linguistic priorities overcome up to the attention the host of bracket fungi. Such a solution superbly reduced in order to movement generated by the Internet and improved the speed and the accuracy of searching.

Polish wersion

Propozycja Priorytet łączenia z host według kraju, języka
Witam. Jestem Polakiem, słabo znam angielski dlatego zamieszczam komputerowe tłumaczenie z polskiego na angielski tekstu z moją propozycją ulepszenia Shareaza.
Proponuję aby można było nadawać priorytety host do których łączy się Shaeraza wg języków w których udostępniamy lub poszukujemy pliki. Ponadto proponuję aby przeszukiwanie host hub również brało pod uwagę priorytety językowe. Takie rozwiązanie znakomicie zmniejszyło by ruch generowany przez internet oraz poprawiło szybkość i trafność wyszukiwań.

So in fact you would like to have people connect to Hubs of the same country to improve searching for files of the same language...?

That is one bad machine translation. I don't think he will be able to understand our answers...

I think he suggests that Shareaza should both prefer uploading to near peers (maybe in a manner similar to how prefer uploaders is done) and prefer search hits from near peers. The first has been suggested before and would probably work, no idea how useful it would be. The second can't really be done with the G2 network topology. Making hubs cluster would indeed improve the performance for files with the specific language, but it would cause serious problems for finding anything else.

First, it would reduce latency, not improve bandwidth- Lower bandwidth can increase efficient bandwidth of a transfer by reducing wait between packets, but query traffic is composed of small discrete packets so it will have pauses between packets irrelevant of connection latency.

If you make hubs prefer connecting to hubs in the same country their connectedness to hubs outside the country is reduced, large countries might in fact become separate islands. And if you don't, just make leaves prefer local hubs, the problem vanishes, but you get mixed clusters and only searching near hubs first becomes impossible. Or would require something like separate caching of near hubs.

I get the idea, but IMHO it is not really compatible with the structure of the Gnutella networks, which relies on high level of interconnectedness between all peers regardless of location.