Hey Everybody,
Has anyone attempted the following.
I was thinking of a hybrid application (windows application) that uses a webserver to communicate through a CGI executable to access a database and do other server related stuff.
Using SocketTools (e.g.) one could call the webserver's CGI exe with parameters e.g. http://mywebserverlocation/CGIAPP.EX...ID&PW=PASSWORD.... (Ok, one should use SSL so the password is not send over in plain-text....)and then do what the application should do. What's the benefit ? Well there is no need to set up a port to be open in your firewall since port 80 is available. The only thing is that you would have to keep track of state, by which I mean that one should log whether the windows application is authorized through it's user ID. One could do this using the
MAC address of the network card but with each request you would need to send it over so maybe a better approach would be that the CGI returns an identifier to the calling application indicating that this user has logged on. Then for each request the CGI should receive this identifier along with whatever you want the CGI do do. The identifier is stored in a field of
a database table.
This is actually some sort of approach to client/server without the need for multi-threading. Now, there is another problem and this problem is logging into a database for each request that'been made. There is something like Connection Pooling but other than with Java I haven't played around with the on a Windows machine.
So, has anybody attempted something like this before. I think this is a pretty neat approach to allow users the benefit from the internet. (e.g. the application could be downloaded from a link on a webpage, unzip the application to a directory and start using it. No installation
necessary plus in the end, when the need falls through to have a web-presence for that application one could still use the same CGI (albeit with different parameters).
Any thought are welcome ?
Steven
Has anyone attempted the following.
I was thinking of a hybrid application (windows application) that uses a webserver to communicate through a CGI executable to access a database and do other server related stuff.
Using SocketTools (e.g.) one could call the webserver's CGI exe with parameters e.g. http://mywebserverlocation/CGIAPP.EX...ID&PW=PASSWORD.... (Ok, one should use SSL so the password is not send over in plain-text....)and then do what the application should do. What's the benefit ? Well there is no need to set up a port to be open in your firewall since port 80 is available. The only thing is that you would have to keep track of state, by which I mean that one should log whether the windows application is authorized through it's user ID. One could do this using the
MAC address of the network card but with each request you would need to send it over so maybe a better approach would be that the CGI returns an identifier to the calling application indicating that this user has logged on. Then for each request the CGI should receive this identifier along with whatever you want the CGI do do. The identifier is stored in a field of
a database table.
This is actually some sort of approach to client/server without the need for multi-threading. Now, there is another problem and this problem is logging into a database for each request that'been made. There is something like Connection Pooling but other than with Java I haven't played around with the on a Windows machine.
So, has anybody attempted something like this before. I think this is a pretty neat approach to allow users the benefit from the internet. (e.g. the application could be downloaded from a link on a webpage, unzip the application to a directory and start using it. No installation
necessary plus in the end, when the need falls through to have a web-presence for that application one could still use the same CGI (albeit with different parameters).
Any thought are welcome ?
Steven
Comment