It can be argued that almost everyone on this list though that
StartServers and MaxServers did define a range (that's what I though).
My understanding now is that they don't.
So now you're telling me that I don't know how my own code works --- a
new height of incivility, even for you.
They do define a range --- not a range on the *total* number of
servers, which is what you mistakenly believed, but rather on the
number of servers which are *free*. If the number of free servers
(i.e., waiting for a request) is fewer than MinFreeServers, more are
created; if it is greater than MaxFreeServers, some of them die off.
That's why I'm calling them MinFreeServers and MaxFreeServers rather
than MinServers and MaxServers. This is about the *simplest*
explanation of the mechanism I can think of.
As to your suggested names, I repeat that they have no mnemonic
significance whatever to anyone who doesn't read the paragraph-long
just-so stories which you made up to justify them. Specifically, as I
have already pointed out, your proposed names (ServersKeptReady and
ServerPoolSize):
*) Don't make it clear that the form of the regulation is to keep the
number of free servers within a range, rather than trying to track
some target value (indeed, the names are likely to mislead people
into thinking that they are target values, which they are *not*).
*) Don't make it clear which is the high side of the range, and which
the low. In fact, I've already forgotten which is which, and
looking at the names, I honestly can't figure it out.
I didn't mention another flaw, BTW; they
*) Don't make it clear that it is the number of *free* servers, rather
than the total number of servers, which is being regulated.
... which is as bad as the other two put together.
Since this whole discussion is about choosing names which will be
suitable for people who don't read the documentation, these problems
render your proposal untenable. -1.
At least with MinFree and MaxFree, they only have to understand one
sentence, which is (all together now):
If the number of free servers (i.e., waiting for a request) is
fewer than MinFreeServers, more are created; if it is greater than
MaxFreeServers, some of them die off.
It helps if they understand another sentence:
This guarantees that servers will always be available for new
clients (MinFreeServers) while also limiting the amount of resources
tied up in idle server processes (MaxFreeServers), rather than
forcing you to guess how many servers you'll need (locking out
clients if you guess too few, and wasting swap space if you guess
too many).
but it's not strictly necessary.
On a personal note --- frankly, Rob, your obtuseness is driving me
crazy. In fact, I'm thinking seriously about leaving it the way it is
right now, and vetoing all changes, because I'm just plain sick and
tired of the whole discussion. If that's not what you want, you may
wish to consider the merit of quitting while you're ahead.
rst
StartServers and MaxServers did define a range (that's what I though).
My understanding now is that they don't.
So now you're telling me that I don't know how my own code works --- a
new height of incivility, even for you.
They do define a range --- not a range on the *total* number of
servers, which is what you mistakenly believed, but rather on the
number of servers which are *free*. If the number of free servers
(i.e., waiting for a request) is fewer than MinFreeServers, more are
created; if it is greater than MaxFreeServers, some of them die off.
That's why I'm calling them MinFreeServers and MaxFreeServers rather
than MinServers and MaxServers. This is about the *simplest*
explanation of the mechanism I can think of.
As to your suggested names, I repeat that they have no mnemonic
significance whatever to anyone who doesn't read the paragraph-long
just-so stories which you made up to justify them. Specifically, as I
have already pointed out, your proposed names (ServersKeptReady and
ServerPoolSize):
*) Don't make it clear that the form of the regulation is to keep the
number of free servers within a range, rather than trying to track
some target value (indeed, the names are likely to mislead people
into thinking that they are target values, which they are *not*).
*) Don't make it clear which is the high side of the range, and which
the low. In fact, I've already forgotten which is which, and
looking at the names, I honestly can't figure it out.
I didn't mention another flaw, BTW; they
*) Don't make it clear that it is the number of *free* servers, rather
than the total number of servers, which is being regulated.
... which is as bad as the other two put together.
Since this whole discussion is about choosing names which will be
suitable for people who don't read the documentation, these problems
render your proposal untenable. -1.
At least with MinFree and MaxFree, they only have to understand one
sentence, which is (all together now):
If the number of free servers (i.e., waiting for a request) is
fewer than MinFreeServers, more are created; if it is greater than
MaxFreeServers, some of them die off.
It helps if they understand another sentence:
This guarantees that servers will always be available for new
clients (MinFreeServers) while also limiting the amount of resources
tied up in idle server processes (MaxFreeServers), rather than
forcing you to guess how many servers you'll need (locking out
clients if you guess too few, and wasting swap space if you guess
too many).
but it's not strictly necessary.
On a personal note --- frankly, Rob, your obtuseness is driving me
crazy. In fact, I'm thinking seriously about leaving it the way it is
right now, and vetoing all changes, because I'm just plain sick and
tired of the whole discussion. If that's not what you want, you may
wish to consider the merit of quitting while you're ahead.
rst