100K

[rejetto]

 

You can read the number of users at this address www.rejetto.com/hfs/?f=awards
I want to say that you'll see the number of users go up and down, not always up, because the counting is made of fresh users (the last month). People who no longer use HFS is just not counted. So that number is real. When you'll visit the page tomorrow, it may be under 100k.

For those of you who wonder how the counting is made: HFS relies on contacting rejetto.com for services like updates, self test, and external IP. What the counter does is to count the number of different IP addresses of such contacts, in the last month. It's updated daily. Since most people has dynamic address, for the A.B.C.D address, the D part is ignored.
Of course this number is not exact, because HFS requires no registration, and I just cannot count the people. But it's  representative, and today it touched the magical 6 digits.

[Foggy]

congrats, and I lol'd at the pic

[Giant Eagle]

for the A.B.C.D address, the D part is ignored.

not smart..

around here (my house + neighbours) we all start with 84.104.178. ---

so basically you count our entire neighbourhood as just 1 user



[ontopic]

Congratulations!!

[TSG]

Yay, I knew this day was coming soon.

For the record www.RAWR-Designs.com has currently supplied around 9000 users, with around 8500 actually visiting our homepage , this is going off download hits and site hits. We are proud to serve the HFS community.

I am unsure how many people have downloaded from our HFS mirror.

[rejetto]

around here (my house + neighbours) we all start with 84.104.178. ---
so basically you count our entire neighbourhood as just 1 user

I know. I guess this is lower esteem, but i guess it is amortised with the people who run HFS just once and decide to not use it.
If you have a smarter method, just suggest.

[MarkV]

Generating a random ID of sufficient length at the start of HFS. The ID is sent on every request and deleted when HFS shuts down. You'd then count only the different IDs, not IPs. Would be a possibility.

[rejetto]

So you start HFS, then restart, and i'd count it twice?

[bacter]

Rejetto, first my congratulations for your successful work.

send with identifier datetime of hfs.exe, then count ip.ip.ip.datetime !

should be a little more accurate (not all people on same ip.ip.ip.xx will install or update at the same time!) , also you don't count hfs working on lans that are not open to wan.

[Giant Eagle]

also you don't count hfs working on lans that are not open to wan.

HFS would be the first if it were able to count its instances running on non-WAN LAN's.. or something


idea:

Create 2 tables in the database called period1 and period2. You store all the ip addresses of the first day into period1.

The next day, you store all the ip addresses in period2. At the end of that day, you search for duplicates in both tables. This result would be the user count. Clear period1 and start filling period1 with the information of the third day.

At the end of that day, search for duplicates and then clear period2, and start re-filling it with the information from the fourth day. etc.. etc..

for more accurate readings (as not all of us leave their fileserver online the whole day) you could make it check every 2 - 3 hours with the same method.

[ELEVENNNN]

This is a truly epic day. and another truly epic waste of a post by me :|

231 total posts, 6 reasonably intelligent ones. YEAA

[rejetto]

send with identifier datetime of hfs.exe, then count ip.ip.ip.datetime !

a guy testing betas would result as 5 different users because of the frequent updates

also you don't count hfs working on lans that are not open to wan.

HFS would be the first if it were able to count its instances running on non-WAN LAN's.. or something

the survey tells only 12% are using HFS in LAN only. I guess that at least an half of these are NATted, and thus counted. I don't think they are a large number.

Create 2 tables in the database called period1 and period2. You store all the ip addresses of the first day into period1.
The next day, you store all the ip addresses in period2. At the end of that day, you search for duplicates in both tables. This result would be the user count. Clear period1 and start filling period1 with the information of the third day.

I don't see why this method should work better. I use HFS few minutes in a week.

[Foggy]

send with identifier datetime of hfs.exe, then count ip.ip.ip.datetime !

how about instead of using the datetime why not get hfs to create a random value on it's first use and then store it in the ini/registry. Then when hfs checks for updates send ip.ip.ip.value

[rejetto]

I fear people would feel tracked. I remember hard feelings when CPUs got a unique ID.

[MarkV]

Why not generating the ID which is valid for, say, 24 hours? Then a new one is generated. This method avoids the restart problem.

[rejetto]

you use today, ID1. you use tomorrow ID2. how would i know you are the same user?
maybe you lost the point on the way