HowManyDonkeys

this page tries to figure out how many donkeys are out there. why? just because.

Contents 1 method 1 instructions 1.1 method 1, discussion: 2 method 2 instructions 2.1 method 2, discussion: 3 Current results by clients

method 1 instructions

i suggest we do this by checking our stats cs or client_stats every once in a while.

if you have _any_ individual score that exceeds* the corresponding one hereunder then edit that number.

• add your name as well, but remember, this is not some sort of competition.

• only do this if you have unstable-2-02-8 or later, since they changed the \"seen\" code.

don't add your info, edit* the current one.

• change just the ones you've exceeded!

---

_please, please, read the instructions first_

uptime: Total Uptime: 635 days, 22h:24m:39s (= 54944679 seconds) (by Santicluke)

total clients seen: 93,425,774 (by DoC)

eDonkeys seen: 12,940,214 (by DoC)

OldMldonkeys seen: 11,712 (by Wyvern)

NewMldonkeys seen: 247,936 (by FtW)

Overnet clients seen: 5,844,641 (by DoC)

Emules seen: 70,302,372 (by DoC)

TrustedMldonkeys seen: 2,431,693 (by DoC)

cDonkeys seen: 3,212 (by DoC)

lMules seen: 10,298 (by censor) (no longer displayed on recent releases)

xMules seen: 126,601 (by DoC)

aMules seen: 506,524 (by Santicluke)

shareazas seen: 1,289,576 (by DoC)

lPhants seen: 1,009,972 (by Santicluke)

VeryCDs seen: 232 (by Santicluke)

Brand			:	Seen	%
eDonkey		:	12940214 (14%)
old mldonkey		:	4708 (0%)
new mldonkey		:	87941 (0%)
Overnet 		:	5844641 (6%)
eMule 			:	70302372 (75%)
trusted mldonkey	:	2431693 (3%)
cDonkey		:	3212 (0%)
xMule 			:	126601 	(0%)
shareaza 		:	1289576 (1%)
aMule 			:	325209	(0%)
lPhant 		:	66374 (0%)
Total 			:	93425774 (100%)
* figures by DoC, running MLDonkey 2.5-30.5 24/7 on a 600kbit link on Apr 23rd 2005

method 1, discussion:

sarefo: are you sure this will be distinct clients, so it's not possible that one client connects again after some hours and is counted again?

_-- hmm, we should probably ask pango how his code works. or rather simon, since he changed it recently /Kokamomi_

_-- sources are identified by their IP address, so you can count several times clients that change of address. /Pango_

Also interesting: [1]

_-- but is this method valid, disregarding the (marginally few?) clients that can change their ip?_

_-- first, I don't think it's so marginal, it seems that many users are disconnected at least once a day, and many get a new IP each time (I do). Secondly, \"seen\" only counts clients that connected to yours, so it counts peers that have a file you're interested in, or that are interested in one of your files. Maybe not what you're expecting..._ 000000

_-- i doubt that this statistic will lead to any valid numbers...but I will try to help you :) /Taros666_

_-- yeah, jigle stats says there's about 750k clients total on the network. Taros666 results are almost there. let's see if we wildly overshoot jigle... and certainly Taros666 can't have files that almost_ everyone _wants =). could anyone think of a better approach? /Kokamomi_

_-- I was only downloading 5 files at the times of my stats. Well, really well spread files but I doubt that the whole network was seen by my mldonkey, because all files where german lang! And as these stats are 5 days old, I think it is the proof, that Method1 will not lead to anything. My latest stat show, that I saw more clients that clients in net. So we might just forget about it? /Taros666_

---

_-- sorry for being obsessed by this =) can we assume that the percentages in our stats are fairly true, if we take the average of several users? maybe by creating a table where folks can input their \"seen\" percentages as column data, one row for each user, with a minimum uptime of 24 hrs or so. like method 2:

method 2 instructions

if you've been on line at least 24 hrs you can enter one (1) line in this table taken from the seens (or successful connections) output of cs or client_stats. later, i'll calculate the averages of the percentages. assuming 750000 clients in total (jigle stats), we might get somewhere. maybe it takes more than 10 people. we'll see =)

• you must have been online at least 24 hrs

• fill in only one row

• you must run unstable-2-02-8 or later

• you should not have slot limitation for eMules, or any other such mechanism, enabled

seen clients:
|-
|user #	||edonkey||old mld||new mld||overnet||new emu||trs mld||cDonkey||lMule  ||uptime ||
+-------+-------+-------+-------+-------+-------+-------+-------+-------+-------||
|-
|1	||12%	||0%	||2%	||7%	||79%	||0%	||0%	||0%	||1+21:17||
|-
|2	||4	||0	||6	||7	||82	||1	||0	||0	||1+10:56||
|-
|3	||3	||0	||7	||6	||84	||1	||0	||0	||1+23:38||
|-
|4	||6	||0	||3	||12	||78	||1	||0	||0	||1+00:29||
|-
|5	||8	||1	||10	||12	||69	||0	||0	||0	||1+13:03||
|-
|6	||13	||0	||5	||10	||71	||0	||0	||0	||1+20:26||
|-
|7	||2	||0	||2	||14	||81	||0	||0	||0	||1+23:16||
|-
|8	||15	||1	||5	||3	||75	||1	||0	||0	||2+16:14||
|-
|9	||12	||0	||3	||16	||68	||1	||0	||0	||1+23:53||
|-
|10	||7	||0	||2	||11	||79	||1	||0	||0	||6+02:29||
|-
|11	||13	||0	||7	||17	||62	||1	||0	||0	||20+21:34||
|-
|12	||19	||0	||4	||11	||65	||1	||0	||0	||6+02:58||
|-
|13	||13	||0	||1	||5	||80	||0	||0	||1	||1+00:02||
|-
|14	||31	||0	||12	||4	||37	||17	||0	||0	||48+19:12||
+-------+-------+-------+-------+-------+-------+-------+-------+-------||
|-
|average||9.5	||0.2	||4.7	||10.5	||74.4	||0.7	||0	||0	||

method 2, discussion:

_-- only 2.02-8 (or more) users should play, because only those know about all the kinds of mldonkey clients; And also, hopefully, they're not yet other clients trying to cheat with that version. Percentages should be computed after excluding Overnet clients, because they aren't eDonkey clients (not part of the \"jigle 750k\"). Or maybe that can be compensated later. All clients filtering (upload slots quotas, large queue banning,... ?) should be removed from the client first, to avoid any slant. The percentage of mldonkeys looks curiously too high, I suppose it's because of slots quotas. Check with platform agnostic files only (music, movies, but not Windows games for example). /Pango_

_-- yup. adding your remarks to instructions. ( don't use slot limitation, since i don't think it's in line with the FileSharing concept). regarding agnostics, i think those effects should be eliminated as more users add their info._

_-- why would slot limitation change something in seen clients ? seeing them doesn't mean you transfer from them !._

_-- Maybe not directly. I don't know. But statistics started being totally unbelievable after the quotas were added. That could be a pure coincidence._

max_emule_slots = 100 is not enough (there's still a reserved slot for MLdonkey), try [2] instead.

Current results by clients

[3]

figure: Total Clients (as seen by jigle.com). Overnet clients are not included in jigle stats.

_a year ago some people didn't want to see more than a 100k users on the network =)_

---

a quick calculation based on the poll on MLdonkey usage (http://mldonkey.berlios.de/modules.php?nameSurveys&pollID13) tells us that about 79% out of the total number of MLdonkey users are online at any one time. current results from method 2 suggests about 6% are MLdonkey users. this means theres about 30.000 MLdonkeys on line. crazy?

---

thanks for your input!

never set \"This is a minor change\" for updating stats on this page. soon enough, updates will be rather few