| |
Main
Date: 11 Aug 2006 11:56:14
From: nospam
Subject: FICS vs icc
|
Things are picking up on FICS! Today I met a handful of fairly strong
players that played interesting chess, and gave me more than I bargained
for. I have noticed more and more players coming over since icc got a
bug up their butt.
Playing on FICS allows you to 'sign on' with an imaginative nick, so
long as it is not registered. I believe this adds some color to the
game, letting you express yourself by witty means. Try it, you'll like it.
Soon, I won't even remember what icc was.
|
|
| |
Date: 16 Aug 2006 12:59:13
From: Thomas T. Veldhouse
Subject: Re: FICS vs icc
|
nospam <[email protected] > wrote:
> Things are picking up on FICS! Today I met a handful of fairly strong
> players that played interesting chess, and gave me more than I bargained
> for. I have noticed more and more players coming over since icc got a
> bug up their butt.
>
> Playing on FICS allows you to 'sign on' with an imaginative nick, so
> long as it is not registered. I believe this adds some color to the
> game, letting you express yourself by witty means. Try it, you'll like it.
>
> Soon, I won't even remember what icc was.
As a registered ICC user who always plays rated games, I have found that there
is MORE opportunity to play since the change. It seems fewer people are put
off by all the guests now. Other than that, I never really cared one way or
the other.
--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
|
| | |
Date: 17 Aug 2006 00:31:08
From: Dave (from the UK)
Subject: Re: FICS vs icc
|
Thomas T. Veldhouse wrote:
> As a registered ICC user who always plays rated games, I have found that there
> is MORE opportunity to play since the change. It seems fewer people are put
> off by all the guests now. Other than that, I never really cared one way or
> the other.
>
Personaly I feel the guests were better than tons of registered users on
7-day trails. The 7-day accounts which are 'throaway' much like hotmail,
gmail and other email addreses.
--
Dave (from the UK)
Please note my email address changes periodically to avoid spam.
It is always of the form: [email protected]
Hitting reply will work for a few months only - later set it manually.
http://witm.sourceforge.net/ (Web based Mathematica front end)
|
| | | |
Date: 18 Aug 2006 12:42:57
From: Thomas T. Veldhouse
Subject: Re: FICS vs icc
|
"Dave (from the UK)" <[email protected] > wrote:
> Thomas T. Veldhouse wrote:
>
>> As a registered ICC user who always plays rated games, I have found that there
>> is MORE opportunity to play since the change. It seems fewer people are put
>> off by all the guests now. Other than that, I never really cared one way or
>> the other.
>>
>
> Personaly I feel the guests were better than tons of registered users on
> 7-day trails. The 7-day accounts which are 'throaway' much like hotmail,
> gmail and other email addreses.
True ... but it required a credit card to sign-up for 7-day accounts, which
limits the ease of using it as a throw-away account. Also, you can always set
a formula to not play unrated players or players that have not yet left the
provisional status.
--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
|
| | | | |
Date: 18 Aug 2006 14:28:32
From: Dave (from the UK)
Subject: Re: FICS vs icc
|
Thomas T. Veldhouse wrote:
> "Dave (from the UK)" <[email protected]> wrote:
>>Personaly I feel the guests were better than tons of registered users on
>>7-day trails. The 7-day accounts which are 'throaway' much like hotmail,
>>gmail and other email addreses.
>
>
> True ... but it required a credit card to sign-up for 7-day accounts, which
> limits the ease of using it as a throw-away account.
No you don't. I've just registered 'testforcc' and no credit card was
needed. In fact, not even a valid email address.
> Also, you can always set
> a formula to not play unrated players or players that have not yet left the
> provisional status.
>
True, but when you see a seek graph, this is not obvious. I clicked on
someone the other day who wanted a 120 0 game, so could last 4 hours.
Then I found I would have been his first rated game. At that point I
decided to not start the game (I said something had cropped up to
honest). I did however point out to him he might find it hard to get 120
0 games until he has an established rating. Even then, 120 0 is going to
be hard to get, but especially as a non-established player.
Although the official line is something along the lines of guest were
using a lot of resources, I find that a bit hard to believe. The amount
of bandwidth required for chess games is very small indeed. I reckon I
could play 100 chess games for the same bandwidth it takes to download
Blitzin or the new Dasher.
Also, a lot of resisted users used guest accounts too. I often play on a
PDA (sometimes in bed!!!) The software has an awful interface, so I'd
not take the games too serious. I'd rather they did not litter my
'recent games' list. I'd play as a guest, and it would be apparent to
anyone easily what I was.
--
Dave (from the UK)
Please note my email address changes periodically to avoid spam.
It is always of the form: [email protected]
Hitting reply will work for a few months only - later set it manually.
http://witm.sourceforge.net/ (Web based Mathematica front end)
|
| | | | | |
Date: 18 Aug 2006 14:15:47
From: Thomas T. Veldhouse
Subject: Re: FICS vs icc
|
"Dave (from the UK)" <[email protected] > wrote:
>
> Also, a lot of resisted users used guest accounts too. I often play on a
> PDA (sometimes in bed!!!) The software has an awful interface, so I'd
> not take the games too serious. I'd rather they did not litter my
> 'recent games' list. I'd play as a guest, and it would be apparent to
> anyone easily what I was.
>
See my previous comment on this. Then read this:
"We've heard the call!
Many paying members have told us that they sometimes like to play
anonymously, but our recent change to guest login doesn't allow them
to do so. We've listened to our members and now plan to add a new
anonymous login feature that will give members the same sort of
access they had in the past as guests. Given our full development
schedule right now, it will likely be several weeks before this can
be implemented, so please bear with us. Thanks for your feedback.
Joel Berez
President"
--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
|
| | | | | |
Date: 18 Aug 2006 14:14:35
From: Thomas T. Veldhouse
Subject: Re: FICS vs icc
|
"Dave (from the UK)" <[email protected] > wrote:
>
> Although the official line is something along the lines of guest were
> using a lot of resources, I find that a bit hard to believe. The amount
> of bandwidth required for chess games is very small indeed. I reckon I
> could play 100 chess games for the same bandwidth it takes to download
> Blitzin or the new Dasher.
>
I am sure bandwidth is not the issue directly. For one, Blitzn' comes from
the web server and not the game servers. Really though, I am sure the
resources issue is related to CPU, and storage [both database and other
types]. The server software is still basically old and I suspect it spawns a
lot of processes [rather than threads] to do its work ... which is bound to be
more resource intensive.
> Also, a lot of resisted users used guest accounts too. I often play on a
> PDA (sometimes in bed!!!) The software has an awful interface, so I'd
> not take the games too serious. I'd rather they did not litter my
> 'recent games' list. I'd play as a guest, and it would be apparent to
> anyone easily what I was.
>
ICC announced that they were going to offer a solution so that registered
users could do something along the lines that you mention [I don't know if it
will be called "guest" or "anonymous" or something else].
--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
|
| | | | | | |
Date: 18 Aug 2006 17:39:48
From: Dave (from the UK)
Subject: Re: FICS vs icc
|
Thomas T. Veldhouse wrote:
> "Dave (from the UK)" <[email protected]> wrote:
>
>>Although the official line is something along the lines of guest were
>>using a lot of resources, I find that a bit hard to believe. The amount
>>of bandwidth required for chess games is very small indeed. I reckon I
>>could play 100 chess games for the same bandwidth it takes to download
>>Blitzin or the new Dasher.
>>
>
>
> I am sure bandwidth is not the issue directly.
Me too.
> For one, Blitzn' comes from
> the web server and not the game servers.
True, although the bandwidth and CPU usage has to be paid for.
> Really though, I am sure the
> resources issue is related to CPU
I really do find it hard to believe that much CPU usage would be used in
doing this. There are typically more computers online that guests. (In
fact, there is normally more computers seeking games than humans seeking
games).
> and storage [both database and other
> types].
But very little storage would be used for guest accounts. And in any
case disks are very cheap now. I know the servers might well use SCSI
disks, which are not the cheapest disks in the world, but you can store
an alful lot of games on a single disk.
The game is not stored beyond the time the person logs out. I assume
once they have gone, that is the end of the record, except in the
account of the paying user.
> The server software is still basically old and I suspect it spawns a
> lot of processes [rather than threads] to do its work ... which is bound to be
> more resource intensive.
It's amazing that FICS seems to manage.
>>Also, a lot of resisted users used guest accounts too. I often play on a
>>PDA (sometimes in bed!!!) The software has an awful interface, so I'd
>>not take the games too serious. I'd rather they did not litter my
>>'recent games' list. I'd play as a guest, and it would be apparent to
>>anyone easily what I was.
>>
>
>
> ICC announced that they were going to offer a solution so that registered
> users could do something along the lines that you mention [I don't know if it
> will be called "guest" or "anonymous" or something else].
Until then, I'll just register a 7-day account for this. Which I'm sure
will take more storage than the guest account, since ICC will have to
store the games.
--
Dave (from the UK)
Please note my email address changes periodically to avoid spam.
It is always of the form: [email protected]
Hitting reply will work for a few months only - later set it manually.
http://witm.sourceforge.net/ (Web based Mathematica front end)
|
| | | | | | | |
Date: 18 Aug 2006 18:00:08
From: Thomas T. Veldhouse
Subject: Re: FICS vs icc
|
"Dave (from the UK)" <[email protected] > wrote:
> Thomas T. Veldhouse wrote:
>
>> For one, Blitzn' comes from
>> the web server and not the game servers.
>
> True, although the bandwidth and CPU usage has to be paid for.
>
Bandwidth is cheap .. VERY CHEAP.
>> Really though, I am sure the
>> resources issue is related to CPU
>
> I really do find it hard to believe that much CPU usage would be used in
> doing this. There are typically more computers online that guests. (In
> fact, there is normally more computers seeking games than humans seeking
> games).
>
A couple of thousand players all on at one time, all consuming one connection
to the server and quite likely a process assigned to each of them. Yes, there
is a significant amount of CPU, I/O, memory associated with that. That ICS
software is old, and it almost certainly is assigning a new process [via a
fork] to each connection.
>> and storage [both database and other
>> types].
>
> But very little storage would be used for guest accounts. And in any
> case disks are very cheap now. I know the servers might well use SCSI
> disks, which are not the cheapest disks in the world, but you can store
> an alful lot of games on a single disk.
>
Any game they play against a rated player gets stored. Further, I am sure
they store information even about guests. Also, there are other resources at
stake.
> The game is not stored beyond the time the person logs out. I assume
> once they have gone, that is the end of the record, except in the
> account of the paying user.
>
>> The server software is still basically old and I suspect it spawns a
>> lot of processes [rather than threads] to do its work ... which is bound to be
>> more resource intensive.
>
> It's amazing that FICS seems to manage.
>
>>>Also, a lot of resisted users used guest accounts too. I often play on a
>>>PDA (sometimes in bed!!!) The software has an awful interface, so I'd
>>>not take the games too serious. I'd rather they did not litter my
>>>'recent games' list. I'd play as a guest, and it would be apparent to
>>>anyone easily what I was.
>>>
>>
>>
>> ICC announced that they were going to offer a solution so that registered
>> users could do something along the lines that you mention [I don't know if it
>> will be called "guest" or "anonymous" or something else].
>
> Until then, I'll just register a 7-day account for this. Which I'm sure
> will take more storage than the guest account, since ICC will have to
> store the games.
>
I wouldn't count on them leaving that hole open for long.
--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
|
| | | | | | | | |
Date: 22 Aug 2006 11:02:39
From: David Richerby
Subject: Re: FICS vs icc
|
Thomas T. Veldhouse <[email protected] > wrote:
> "Dave (from the UK)" wrote:
>> Thomas T. Veldhouse wrote:
>>> Really though, I am sure the resources issue is related to CPU
>>
>> I really do find it hard to believe that much CPU usage would be
>> used in doing this. [...]
>
> A couple of thousand players all on at one time, all consuming one
> connection to the server and quite likely a process assigned to each
> of them. Yes, there is a significant amount of CPU, I/O, memory
> associated with that. That ICS software is old, and it almost
> certainly is assigning a new process [via a fork] to each
> connection.
These processes will spend almost all their time blocked waiting for
I/O: the CPU usage of even a few thousand of those isn't all that
great. Most of the memory used by the processes (i.e., the code
itself) will be shared: they don't need an awful lot of state so they
don't need much memory each. Shouldn't be a problem.
>>> and storage [both database and other types].
>>
>> But very little storage would be used for guest accounts.
>
> Any game they play against a rated player gets stored.
Straw man: the registered player's games will get stored whether they
are against guests or not. Further, if there are large numbers of
games between registered users and guests, doesn't that indicate that
the registered users want to pay for the service of paying against
guests?
Dave.
--
David Richerby Confusing Laptop Priest (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a man of the cloth that you
can put on your lap but you can't
understand it!
|
| | | | | | | | | |
Date: 22 Aug 2006 12:58:31
From: Thomas T. Veldhouse
Subject: Re: FICS vs icc
|
David Richerby <[email protected] > wrote:
> Thomas T. Veldhouse <[email protected]> wrote:
>> "Dave (from the UK)" wrote:
>>> Thomas T. Veldhouse wrote:
>>>> Really though, I am sure the resources issue is related to CPU
>>>
>>> I really do find it hard to believe that much CPU usage would be
>>> used in doing this. [...]
>>
>> A couple of thousand players all on at one time, all consuming one
>> connection to the server and quite likely a process assigned to each
>> of them. Yes, there is a significant amount of CPU, I/O, memory
>> associated with that. That ICS software is old, and it almost
>> certainly is assigning a new process [via a fork] to each
>> connection.
>
> These processes will spend almost all their time blocked waiting for
> I/O: the CPU usage of even a few thousand of those isn't all that
> great. Most of the memory used by the processes (i.e., the code
> itself) will be shared: they don't need an awful lot of state so they
> don't need much memory each. Shouldn't be a problem.
>
Each process gets a copy of the entire environment, unlike a threaded service
which shares one. Process forks are expensive. Unless you know what machines
they run under what current load, you can not possibly judge whether resources
are a valid concern or not. What I can say is that it appears they are
running a branch of the old ICS software that FICS also runs, and that is old
code based upon process forking. They are much more likely to be hitting the
resource limits with that code than they are with perhaps a more modern
implementation.
--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
|
|
