status of DHT feature?

[BlackLion]

It's been said many times that the devs are going to support mainline/BitComet DHT. In the future, it might be a possibility that they'll add Azureus DHT also.

That would definitely make it the client of choice, either one would but both would definitely. One can only wish right?

[Veneficus]

Considering how hard it is to get one implementation right due to a lack of docs, if the Azureus one is ever implemented, it will be a LONG time from now.

[Firon]

Well, ludde's smart, he's already 95% done with the mainline DHT implementation.

I imagine it'd be much easier to implement the Azureus one since he'd already know the basics of DHT?

[Xilon]

Good to hear about the progress of DHT. I don't really like the idea of having a non-standard (Azureus) implementation of this feature though. Programs should stick to standards so that you don't end up like the eMule mods that have 50 different credit systems or something. Standards are there for a reason

[JigPu]

The reason DHT implementation is going slowly is that it is not documented in any way. Even the source code in the mainline client is cryptic at best, and contains no comments. We went as far as reverse engineering the protocol through packet sniffing to find out exactly what was going on under the hood and used the python source to confirm our findings. I think we made a substantial amount of progress today, and I believe it will be usable soon. Most likely before 1.2 comes.

Ouch. I feel your pain. A couple weeks ago I had to do the same thing for a different program, and it certianly isn't any fun. Good to hear that it's still coming along though

/me goes on a rampage against those who can't comment code/document features

JigPu

[BlackLion]

Good to hear about the progress of DHT. I don't really like the idea of having a non-standard (Azureus) implementation of this feature though. Programs should stick to standards so that you don't end up like the eMule mods that have 50 different credit systems or something. Standards are there for a reason

agreed.

[ReP0]

Good to hear about the progress of DHT. I don't really like the idea of having a non-standard (Azureus) implementation of this feature though. Programs should stick to standards so that you don't end up like the eMule mods that have 50 different credit systems or something. Standards are there for a reason

agreed.

I generally agree but the creator of the mainline dht implementation dropped the ball. Just look at how much trouble utorrent is having implementing this so called standard. I don't blame azureus for going ahead and saying f@#$ it and doing their own.

A standard is good to follow as long as it's available. Reverse engineering some silly ass protocol and calling it some standard by sniffing packets or trying to decipher code source is pathetic. If I was doing utorrent I'd go azureus implemetation just to teach the mainline bastards a lesson to write some documentation and given azureus is very common at the moment I'd win. If and when it came out officially I would think about implementing it. Very much the attitude a few other clients are adopting. Utorrent coders are being too nice I think or at least playing a stategic game of isolating azureus users and splitting them from the herd.

[ASDF Commander]

Who's to say Azureus DHT is non-standard? They created the standard (they were the first, sorry), and Az DHT has the most users. I don't understand all the animosity towards Az in this forum. It's one of, if not, the fastest client with regard to xfer speeds, it's free, it's open source, it's supported on most platforms and it's not like they're forcing you or locking you into their client. If you don't like it, move onto to something else, but no need to constantly bash it.

[Keloran]

reson it gets bashed is because most of us have tried it and hate it, if we didnt hate it we would have no reason to bash it

[ASDF Commander]

I wasn't denying anyone hated it, I was asking why. I understand it's slugish, but in every other respect it's superior.

[chaosblade]

"It's one of, if not, the fastest client with regard to xfer speeds"

Very very user-subjective.

"it's free, it's open source"

So what? that doesnt say anything about its quality.

"it's supported on most platforms"

Again, doesnt tell you anything about its performance as a torrent client.

"and it's not like they're forcing you or locking you into their client. If you don't like it, move onto to something else, but no need to constantly bash it."

Hence the reason we are here. Why do you feel the need to 'defend' Azureus ?

[1c3d0g]

@ ASDF Commander: Didn't you notice? You poor bastard. :/ Azureus:

● sucks dry ~20% of your CPU

● gobbles up 200+ MB of RAM

● bloated with useless features

µTorrent tries to avoid all these issues, and so far has accomplished this wonderfully. µTorrent is like a breath of fresh air in a world where poison is saturating the air. That's why everyone here likes µTorrent, and so should you. :cool:

[ASDF Commander]

"It's one of, if not, the fastest client with regard to xfer speeds"

Very very user-subjective.

"it's free, it's open source"

So what? that doesnt say anything about its quality.

"it's supported on most platforms"

Again, doesnt tell you anything about its performance as a torrent client.

"and it's not like they're forcing you or locking you into their client. If you don't like it, move onto to something else, but no need to constantly bash it."

Hence the reason we are here. Why do you feel the need to 'defend' Azureus ?

I wouldn't say it's subjective. You have a data rate, theres nothing subjective about that. Torrent speed, however, is very much dependent on the individual torrent and luck with connections. So, yeah, it is quite hard to judge. The only things I can say that have any real value as data are that I am consistently able to max my connection using Azurues and also that Azurues is very quick to start uploading at full capacity unlike utorrent (in my experience thus far).

Support for most platforms is just a nice feature.

I defend it because you're bashing a client which IMHO is THE best baring its poor resource usage. I think a lot of you over exaggerate its bloatedness though. Usually only 0-2% CPU usage <80mb of mem for me. That may seem like a lot but I do have a fair ammout of memory so I don't notice.

Just because I'm defending Az doesn't mean I don't like utorrent. I like it a lot, it has a lot of potential. I'm just not switching to it 100% right now because of reasons I have already stated.

[asdfsauce]

@ ASDF Commander: Didn't you notice? You poor bastard. :/ Azureus:

● sucks dry ~20% of your CPU

● gobbles up 200+ MB of RAM

● bloated with useless features

captain exaggeration?

not even close to the real resource usage.

[Animorc]

Let's just say that Azureus is not as good as µTorrent.

Azureus uses somewhere between 40 and 200 MB of the memory. µTorrent isn't even close to the smaller of them.

Azureus installer is 6.7M MB. µTorrent is incredibly far away from that.

And noone that has used µTorrent can say that Azureus is not bloated.

[Veneficus]

This thread is rapidly devolving into an Azureus vs. µTorrent debate. Stop it. We have enough of those on the board I'm sure. Let's get back to DHT.

I don't think it would be noticeably easier to implement Azureus DHT after learning the mainline DHT protocol because they're, well, different. I don't know how different they are. It could be a small difference or a large one, such as the OSCAR protocol versus MSNP9. Both are used for chatting, but they're radically different.

[Keloran]

by the looks of the tests earlier ludde has got it working, not sure which version of DHT though

[Sanchez]

If this goes any further into Az vs utorrent its getting locked.

[1c3d0g]

@ Keloran: Mainline's version, as it should.

[chaosblade]

And after i learned the az dev team are going to implement the mainline method in their client aswell, there's no reason what so ever to implement the az-only one.

[BlackLion]

And after i learned the az dev team are going to implement the mainline method in their client aswell, there's no reason what so ever to implement the az-only one.

Well that settles it.

[ScubaSteve]

i just hope that after all this is implemented someone writes out proper documentation for it, has anyone actually tested which DHT is more efficient. whether it be mainline or azureus, because i absolutely love the azureus DHT.

[Firon]

Where did you see that, chaosblade? I'd like to know where. Link would be nice

[1c3d0g]

Where did you see that, chaosblade? I'd like to know where. Link would be nice

Well actually that was announced in the IRC channel by ludde I think... :/

[BlackLion]

i just hope that after all this is implemented someone writes out proper documentation for it, has anyone actually tested which DHT is more efficient. whether it be mainline or azureus, because i absolutely love the azureus DHT.

Well the Comet's implementation aint exactly chopped liver, they both have gotten me files I would not have gotten otherwise. Just bring the soup when its ready I say