Scratch

flashgocrazy wrote:

here is about mesh

lets make a world wide mesh system that's easier to use and install.

developers
flashgocrazy Director
we need..
a html designer
a php designer
a squeak/blocks designer
and somebody to write tutorials how to use/install
apply to be a dev in a reply!

it would have ....

password protected wwms
php codding
and more!

if your a dev, post your code/tuts for wwm
down there!
|
|
|
U

Ever heard of Scratch Live? Its in claw, it basically posts variables online.

If you want a WWM then just port-forward port 42001.

http://www.scratch.mit.edu/ext/youtube/?v=Kp-R-eHiQco Instead 51329, use 42001. That should be most people's routers but, if it is not, the concept is the same: Click "Applications & Gaming" or "Port-forwarding" then fill in the needed info. Sometimes there is an extra option you need to look/click before you can click either of those like "Advanced Setup" then "Port-forwarding".

Did you just completely ignore Magnie's post there? I think he had a good point - just use port forwarding! 

Also, did you actually use that php code I wrote for you? I spent about three hours on that and never even heard back from you whether you used it or what?

flashgocrazy wrote:

sparks wrote:

Did you just completely ignore Magnie's post there? I think he had a good point - just use port forwarding! 

Also, did you actually use that php code I wrote for you? I spent about three hours on that and never even heard back from you whether you used it or what?

it works

Oh good 

I have an idea...I'm currently writing a program in C++ which will connect to Scratch and (possibly) directly to a remote MySQL host. This would cut down on server-side CPU usage  and MySQL connections. Here's a diagram...

Last edited by ohaiderstudios (2011-11-28 20:07:44)

Two words for you: HTTPSocket and GET. Scratch can read web-pages into variables using its HTTP socket. GET allows a PHP script to pull values from an URL to decide what to do (That's how my self-updating images and links works).

scratch wrote:

(read contents of page at URL [http://www.scratchWWM.com/connect.php?action=post&user=sparks&text=hello_world])

Scratch "views the webpage, triggering the PHP script on it.

PHP wrote:

<?php
$username = $_GET['user'];
if($_GET['action'] == "post"){
     $message = $_GET['text'];
     mysql_query("INSERT INTO `wwm` yadayadayada");
}
?>

can't remember the proper MySQL insert query syntax off the top of my head but you get the idea. This way you don't need an external program. Think about it, it's fine to have a C++ project doing the connecting, but the user then has to (work out how to) start up and run that project as well as the Scratch project.

ohaiderstudios wrote:

I have an idea...I'm currently writing a program in C++ which will connect to Scratch and (possibly) directly to a remote MySQL host. This would cut down on server-side CPU usage  and MySQL connections. Here's a diagram...
http://imageflock.com/img/1322528810.png

awesome idea!

Magnie wrote:

sparks wrote:

Two words for you: HTTPSocket and GET. Scratch can read web-pages into variables using its HTTP socket. GET allows a PHP script to pull values from an URL to decide what to do (That's how my self-updating images and links works).

scratch wrote:

(read contents of page at URL [http://www.scratchWWM.com/connect.php?action=post&user=sparks&text=hello_world])

Scratch "views the webpage, triggering the PHP script on it.

PHP wrote:

<?php
$username = $_GET['user'];
if($_GET['action'] == "post"){
     $message = $_GET['text'];
     mysql_query("INSERT INTO `wwm` yadayadayada");
}
?>

can't remember the proper MySQL insert query syntax off the top of my head but you get the idea. This way you don't need an external program. Think about it, it's fine to have a C++ project doing the connecting, but the user then has to (work out how to) start up and run that project as well as the Scratch project.

The problem is, it takes time ( about 0.2-1 seconds, and 0.05 seconds in Python, which is a *ton* of time, while most other commands run at 0.001 seconds ) to use HTTP requests. And if someone expects to make a game with moving players around with x and y positions, it will be very slow.

If you also so doing it through C++ to connect to a database might/should be slightly faster, but Remote Connections also has the slow "sending a request, receiving a reply". Which is the main problem, if we can cut out all the extra time on the Remote Connections, then Remote Connections will be the most reliable for games. Even now, it is probably the most reliable, though it is still slow and since it is receiving data slower than anything it is connected to, it records all the data sent to it and then sends it to the actual Scratch Project which makes it so nothing is in real-time, so if someone is playing a Multiplayer game, they will fall behind on what is happening, while players with faster computers and internet connections are closer to what is really happening ( but not in real-time ).

If it is possible to just remove the entire wait-time between each block, or just remove the backlog on Remote Connections, it will speed things up by a second or so.

Going back to the HTTP requests, another reason why it takes so long, is because there is a ton of info that is received that is mostly used for telling a browser how to display the text ( like the HTTP Header, Content-Type, then HTML if there is any ). So if you can remove the extra stuff that is sent and received by the HTTP request then you could probably speed up communication by 0.1 seconds ( maybe more? ) and it would use less bandwidth.

Though I'm not sure how that would be done, especially for free-hosts.

What are your thoughts?

See, instead of using PHP to GET requests from a database, which would take up a ton of bandwidth and CPU power, a direct connection to MySQL through C++ would be cleaner and faster. That means NO PHP mediation.

G0D_M0D3 wrote:

I can help with instructions and even make a page on my new website about it when it's done. And I can advertise when I get Scratcher status. And I can make a better banner for it.

ok hired for artist and instruction maker

G0D_M0D3 wrote:

flashgocrazy wrote:

G0D_M0D3 wrote:

I can help with instructions and even make a page on my new website about it when it's done. And I can advertise when I get Scratcher status. And I can make a better banner for it.

ok hired for artist and instruction maker

Does this have a name yet? I was thinking about making the logo like this:

_____________________________________________
'placeholder_name'

"Let's make a new world-wide-mesh"
_____________________________________________

it has a name its called wordzord

ohaiderstudios wrote:

See, instead of using PHP to GET requests from a database, which would take up a ton of bandwidth and CPU power, a direct connection to MySQL through C++ would be cleaner and faster. That means NO PHP mediation.

Yes, that is true. Though there is still the general slowness of the connection/Scratch's Remote Connections. If you do it that way, you will only be accessing the MySQL server. The C++ coding just seems to be the link between Scratch and the MySQL server. So it will be insecure, and people will have full control over the data in there. If you have the C++ connect to a server ( that could also be written in C++ ), 1) everything just jumps a level of security 2) you can create services on there, like the Scratch Live, and maybe even something with the Scratch API and Sparks' API 3) It in general just adds a lot more functionality, and maybe even more flexibility.

And if you really want to get serious, you could code the C++ link to the server, into Squeak instead. No need to run extra programs. ( That was my problem when making my Python Mirror project )

I have tried making a multiplayer game with Python and Remote Connections, and when I press the up-arrow key twenty times and it is broadcasting to the server that that key is being pressed, then the server sending a message back, all the info sent from the server to the client is put into a line and then first data in line is run and it does that for 5-10 seconds after I finish pressing the up-arrow key. I'm thinking that if that 'line-up' is taken out of Scratch/Squeak and have everything running at real-time rather than have each command take it's turn, so you can never really see what is actually going on.

While using Mesh and creating my racing game ( with up to four players ), one is on Ubuntu with low RAM ( 512MB I think, maybe less ) and my Windows 7 Starter which has 1GB of RAM and probably better speed in general. While I am on my Windows 7 Starter, playing the game and having a sibling playing on the Ubuntu, I can get around the entire track on my screen before my sibling even sees me half-way around the track on their screen. What's happening is the 'line-up' I've been talking about, one computer sends 10 messages to the other computer and the messages get in line and wait their turn.

(A) Have you ever typed in a word document and when the process becomes slow and you stop typing, but the letters you were typing continue typing because they are behind? (B) And with another word program, you type and when the process slows down, it stops typing after you have stopped typing though it may not have gotten any other keys you may have typed?

(A) is good for things that must not be skipped. But (B) is better for those games where if it slows down, may as well just put the player in the correct spot when the process speeds up again, or for slow connections in general, may as well just use (B) for games.

I'll just repeat what I had said up there in easier readability terms:
(A) is good for chat, text stuff, things that are used right now, like with Scratch Live.
(B) is good for the games and stuff with graphics and constantly updating variables.


Also, an idea for people who are making programs with servers ( and stuff like your project ohaiderstudios ) instead of having the client check for updates every 0.5 seconds, have the server send out the updates to the clients, so the client is waiting instead of the server. It uses less bandwidth over time.

That is one of the problems with chat programs these days, is they always ask the server for stuff, the server should just send the stuff that the client wants without needing to continually check for it.

Not to be rude or anything, but that is the problem with Scratch Live and iGlobalChat. It will consume so much bandwidth ( which is why web hosts usually ban chat programs ) that the ISP or host will shut down because of over-use. So Scratch Live needs to check with their host's ToS and add those restrictions to their own ToS as well. That goes for anyone else that plans on setting up their own SL.

If Scratch Live were a server where the client asks to be sent any updates on certain variables, that when another client updates the variable, the server sends all those that are "following" that variable, the updated content of the variable name.

I would actually be willing to help (every time I "help" with coding something, I end up doing it myself    ) code the server for something like that, but the only problem is getting a host for it. So those people out there hosting Minecraft servers and advertising here on Scratch, mind lending some space and a port? 

Note: Everything below the first two paragraphs are just more thoughts I decided to throw out. The first two paragraphs were to answer the main quote at top. 

Edit: Wow, that is a lot of text... Hopefully some of it has some good thoughts in it.

Last edited by Magnie (2011-11-30 00:22:04)

Magnie wrote:

ohaiderstudios wrote:

See, instead of using PHP to GET requests from a database, which would take up a ton of bandwidth and CPU power, a direct connection to MySQL through C++ would be cleaner and faster. That means NO PHP mediation.

Yes, that is true. Though there is still the general slowness of the connection/Scratch's Remote Connections. If you do it that way, you will only be accessing the MySQL server. The C++ coding just seems to be the link between Scratch and the MySQL server. So it will be insecure, and people will have full control over the data in there. If you have the C++ connect to a server ( that could also be written in C++ ), 1) everything just jumps a level of security 2) you can create services on there, like the Scratch Live, and maybe even something with the Scratch API and Sparks' API 3) It in general just adds a lot more functionality, and maybe even more flexibility.

And if you really want to get serious, you could code the C++ link to the server, into Squeak instead. No need to run extra programs. ( That was my problem when making my Python Mirror project )

I have tried making a multiplayer game with Python and Remote Connections, and when I press the up-arrow key twenty times and it is broadcasting to the server that that key is being pressed, then the server sending a message back, all the info sent from the server to the client is put into a line and then first data in line is run and it does that for 5-10 seconds after I finish pressing the up-arrow key. I'm thinking that if that 'line-up' is taken out of Scratch/Squeak and have everything running at real-time rather than have each command take it's turn, so you can never really see what is actually going on.

While using Mesh and creating my racing game ( with up to four players ), one is on Ubuntu with low RAM ( 512MB I think, maybe less ) and my Windows 7 Starter which has 1GB of RAM and probably better speed in general. While I am on my Windows 7 Starter, playing the game and having a sibling playing on the Ubuntu, I can get around the entire track on my screen before my sibling even sees me half-way around the track on their screen. What's happening is the 'line-up' I've been talking about, one computer sends 10 messages to the other computer and the messages get in line and wait their turn.

(A) Have you ever typed in a word document and when the process becomes slow and you stop typing, but the letters you were typing continue typing because they are behind? (B) And with another word program, you type and when the process slows down, it stops typing after you have stopped typing though it may not have gotten any other keys you may have typed?

(A) is good for things that must not be skipped. But (B) is better for those games where if it slows down, may as well just put the player in the correct spot when the process speeds up again, or for slow connections in general, may as well just use (B) for games.

I'll just repeat what I had said up there in easier readability terms:
(A) is good for chat, text stuff, things that are used right now, like with Scratch Live.
(B) is good for the games and stuff with graphics and constantly updating variables.


Also, an idea for people who are making programs with servers ( and stuff like your project ohaiderstudios ) instead of having the client check for updates every 0.5 seconds, have the server send out the updates to the clients, so the client is waiting instead of the server. It uses less bandwidth over time.

That is one of the problems with chat programs these days, is they always ask the server for stuff, the server should just send the stuff that the client wants without needing to continually check for it.

Not to be rude or anything, but that is the problem with Scratch Live and iGlobalChat. It will consume so much bandwidth ( which is why web hosts usually ban chat programs ) that the ISP or host will shut down because of over-use. So Scratch Live needs to check with their host's ToS and add those restrictions to their own ToS as well. That goes for anyone else that plans on setting up their own SL.

If Scratch Live were a server where the client asks to be sent any updates on certain variables, that when another client updates the variable, the server sends all those that are "following" that variable, the updated content of the variable name.

I would actually be willing to help (every time I "help" with coding something, I end up doing it myself    ) code the server for something like that, but the only problem is getting a host for it. So those people out there hosting Minecraft servers and advertising here on Scratch, mind lending some space and a port? 

Note: Everything below the first two paragraphs are just more thoughts I decided to throw out. The first two paragraphs were to answer the main quote at top. 

Edit: Wow, that is a lot of text... Hopefully some of it has some good thoughts in it.

I think I've got what you're saying. 
I'm having a horrible time learning winsock, so that might not be done for a while...
But yeah, one thing that I really agree with is:

Magnie wrote:

That is one of the problems with chat programs these days, is they always ask the server for stuff, the server should just send the stuff that the client wants without needing to continually check for it.

I'll just have them make a "when I receive 'the_server_is_sending_data'" script which updates the variables to their sensor counterparts. XD