sockets

[Kenneth Brody]

On 5/18/2012 1:08 PM, Ed Hilovsky wrote:
> Therein lies the problem. FP has "designated" a new way to count socket
> connections.

That's just it...  We're *not* counting "socket connections".

> If there is a socket server and I connect to it, it shows one
> connection. When I do a socketclose it shows 0 connections.

If there's a "socket server", then there's 1 "socket server".  If you
connect to it 50 times, there's still 1 "socket server".

> Whatever logic you are using is beyond me. Bottom line, look at what
> people are saying. The majority say FP is wrong in the way they count
> socket connections.

You keep making the argument "if other programs can count apples correctly,
why can't filePro count them correctly", and my answer has consistently been
"we don't count apples, we count apple baskets, and we count them just fine".

> Once I open a socket in filepro, you count that process that
> connect(ed).

No, it doesn't, as it doesn't count connections.

> Then even if I don't have an active socket open and
> connected, you are counting it against the license. You say this as
> though all of us should just understand that the socket license was
> always sold linked to fp processes.

Let's say you buy a 16-user runtime license.  You have a process that uses 
the main file and 5 others via lookups.  Do you use 1 runtime license or 6?

> Well, I never saw that. I was never
> give that interpretation. I was just told the price and that I needed a
> development socket license also. So, without a firm disclaimer about how
> FP counts, I use what the rest of the world does. And the rest of the
> world doesn't link socket connections to anything other than connect and
> socketclose. It's simple, connect to a socket at a specific port and IP
> and when done do a socketclose. Period...............

If we were licensing connections, I'd agree with you.  But, that's not what 
is being counted.

-- 
Kenneth Brody