Generic Dashboards?

Status
Not open for further replies.
CCarter said:
I'm still going to release it open source, I'll charge for customization, and help setting up. Anyone will be able to develop on it. I just finish 2 beta tester's setups, i'm going down the list guys for betas, so don't worry guys!!
Click to expand...

fuck that. you're going to deal with so many morons that route. make this a SaaS and charge a reasonably high amount for it. you'll get fewer questions or complaints b/c people who pay a premium for this likely won't be retarded.

grandfather in the beta testers (i'm not one) and then charge for it.
 


Medici said:
fuck that. you're going to deal with so many morons that route. make this a SaaS and charge a reasonably high amount for it. you'll get fewer questions or complaints b/c people who pay a premium for this likely won't be retarded.

grandfather in the beta testers (i'm not one) and then charge for it.
Click to expand...

very much this. individual installs of this are pointless, just make it use some sort of plugin architecture and host it for all the users yourself.
 
Medici said:
fuck that. you're going to deal with so many morons that route. make this a SaaS and charge a reasonably high amount for it. you'll get fewer questions or complaints b/c people who pay a premium for this likely won't be retarded.

grandfather in the beta testers (i'm not one) and then charge for it.
Click to expand...

Here is the problem, most Affiliate, hacker, black haters, web developers are already paranoid with people tracking them, so if create a SaaS with this, the adoption level is going to be pretty low. Imagine, having all your affiliate logins, passwords, stats all in one location, on SOMEONE ELSE's server? Not likely going to happen.

Imagine a guy doing serious numbers in a niche, and wanting to use this server, but not wanting to give away his exposure, even if only the trusted administrators are the only ones that can see shit. That's never going to work, and it will completely kill the opportunity. What I was thinking though, was have the open source version still available, but also have a hosted version, that people can sign up for if they don't want the headaches of hosting it themselves, similar to Wordpress.

Also, all OFFICIAL customization is going to go through a reliable web development team which I will get a kick back, but anyone will be able to develop for the system. - eliminating my problem of peasants trying to contact me directly about their particular configurations. As a web developer, I'm already not cheap, so trust me, I've been down similar roads before.

If I release it as a simple SaaS service, it'll be like the other dozen UN-reliable and UN-customizable dashboard already available. All plugins and shit would have to be 100% supported, and then I'll really be dealing with morons, plus trying to help the affiliate community falls by the wayside to making money. I plan on making money indirectly once I get the project's exposure and adoption rate through the roof, which I get with open source. Yeah, I know Saas, I'll be making money, but I'll really don't want to deal with your bullshit. I'll making the platform, plugins, and send it out into the open source world. People will modify it, making it better, etc.
 
CCarter said:
Here is the problem, most Affiliate, hacker, black haters, web developers are already paranoid with people tracking them, so if create a SaaS with this, the adoption level is going to be pretty low. Imagine, having all your affiliate logins, passwords, stats all in one location, on SOMEONE ELSE's server? Not likely going to happen.

Imagine a guy doing serious numbers in a niche, and wanting to use this server, but not wanting to give away his exposure, even if only the trusted administrators are the only ones that can see shit. That's never going to work, and it will completely kill the opportunity. What I was thinking though, was have the open source version still available, but also have a hosted version, that people can sign up for if they don't want the headaches of hosting it themselves, similar to Wordpress.

Also, all OFFICIAL customization is going to go through a reliable web development team which I will get a kick back, but anyone will be able to develop for the system. - eliminating my problem of peasants trying to contact me directly about their particular configurations. As a web developer, I'm already not cheap, so trust me, I've been down similar roads before.

If I release it as a simple SaaS service, it'll be like the other dozen UN-reliable and UN-customizable dashboard already available. All plugins and shit would have to be 100% supported, and then I'll really be dealing with morons, plus trying to help the affiliate community falls by the wayside to making money. I plan on making money indirectly once I get the project's exposure and adoption rate through the roof, which I get with open source. Yeah, I know Saas, I'll be making money, but I'll really don't want to deal with your bullshit. I'll making the platform, plugins, and send it out into the open source world. People will modify it, making it better, etc.
Click to expand...

Independents will present that problem but many agencies won't. You need only a small % of the market to make a very reasonable return on this.

I understand your need to make this customizable and wide reaching but it just seems like too much money to pass up. either way, it looks great and I can't wait to play with it.
 
I have done a windows C program which pulls me adsense, cpmstar, cpalead and such, generating a small html file from it, all 5 minutes updated.
There i can see my income and such.

Then i show the data by using windows gadget!
I think i have with that the one and only useful app for windows gadgets.

Adsense is a bitch to pull because some strange and often changing redirects when you login.
Google seems to work on that part every free minute. They do some work
to make one big login space for all their services.

Pull the mobile version, thats alot cleaner.

Clickbank sucks to pull too, their statistic pages gives even a bot headaches.
 
How are you pulling clickbank? It is simple to do it with a live robot that logins for you through a browser cause it can interact with javascript, but making it a script, text bot is more difficult since it's all javascript for the login, and creating a robot for that which will be able to run out of the box for the masses is without a thousand and one libraries is the challenge. I definitely want to trade notes if you pulled clickbank off.
 
CCarter said:
How are you pulling clickbank? It is simple to do it with a live robot that logins for you through a browser cause it can interact with javascript, but making it a script, text bot is more difficult since it's all javascript for the login, and creating a robot for that which will be able to run out of the box for the masses is without a thousand and one libraries is the challenge. I definitely want to trade notes if you pulled clickbank off.
Click to expand...

Well, i use C with curl (libcurl) - so its hardcore (note the missing ++ by the C).

I call curl like this
sprintf(posttag, "nick=%s&pass=%s&login=Log+In&rememberMe=true&j_username=%s&j_password=%s", CB_LOGIN, CB_PASS,CB_LOGIN, CB_PASS);
sprintf(dtemp, "https://%s.accounts.clickbank.com/account/login", CB_LOGIN);

Then nothing special, i just feed curl in post and https mode. Works fine.

Jsut a note: You need to "urliaze" your password/nick before you put it in curl, curl will not do it for you - and clickbank will fail when your password
includes special chars which a browser will automatically transform. I forgot that last time after worked weeks with a web tool. And i wondered why
it did not worked.

Clickbank will answer with a 302 to your account page. From there the fun start, because the stats page need a hell of regex to get good data.

I suggest to use wget or curl from a (bash) script - it should work too.
 
markbush said:
Well, i use C with curl (libcurl) - so its hardcore (note the missing ++ by the C).

I call curl like this
sprintf(posttag, "nick=%s&pass=%s&login=Log+In&rememberMe=true&j_username=%s&j_password=%s", CB_LOGIN, CB_PASS,CB_LOGIN, CB_PASS);
sprintf(dtemp, "https://%s.accounts.clickbank.com/account/login", CB_LOGIN);

Then nothing special, i just feed curl in post and https mode. Works fine.

Jsut a note: You need to "urliaze" your password/nick before you put it in curl, curl will not do it for you - and clickbank will fail when your password
includes special chars which a browser will automatically transform. I forgot that last time after worked weeks with a web tool. And i wondered why
it did not worked.

Clickbank will answer with a 302 to your account page. From there the fun start, because the stats page need a hell of regex to get good data.

I suggest to use wget or curl from a (bash) script - it should work too.
Click to expand...

urgh, this is so 90s.

https://github.com/kennethreitz/requests
 
mattseh said:
urgh, this is so 90s.

https://github.com/kennethreitz/requests
Click to expand...

Yeah, but the result is a system exe from 18kb which pulls and works out more a half a dozen aff sites. It would work even on your 80s casio, i assume.

I can install it as service on my ultrabook and it use the resources in the background.

People today have no idea how fast their computers really are.
With Python and such you waste 80% of the possibilities. I mean it.

I use for my real projects mainly pure linux bash script and the linux tool set.
I can do things with that on a $9 vserver where others here have to buy whole suites and a dedi.
 
markbush said:
Yeah, but the result is a system exe from 18kb which pulls and works out more a half a dozen aff sites. It would work even on your 80s casio, i assume.

I can install it as service on my ultrabook and it use the resources in the background.

People today have no idea how fast their computers really are.
With Python and such you waste 80% of the possibilities. I mean it.

I use for my real projects mainly pure linux bash script and the linux tool set.
I can do things with that on a $9 vserver where others here have to buy whole suites and a dedi.
Click to expand...

Server costs are miniscule compared to my income, having several servers simply doesn't matter compared to my time. Even then, bandwidth, not CPU / memory is usually my limitation. YMMV.
 
markbush said:
Yeah, but the result is a system exe from 18kb which pulls and works out more a half a dozen aff sites. It would work even on your 80s casio, i assume.

I can install it as service on my ultrabook and it use the resources in the background.

People today have no idea how fast their computers really are.
With Python and such you waste 80% of the possibilities. I mean it.

I use for my real projects mainly pure linux bash script and the linux tool set.
I can do things with that on a $9 vserver where others here have to buy whole suites and a dedi.
Click to expand...

just as a general application theory sidenote: If something is running in the background, speed is much less of an issue. As your app is just polling every 5 minutes, it doesn't matter if it takes 5 seconds or even up to 5 minutes to process.

Doesn't negate your point, but it is an important thing to remember. Background processes allow a lot of flexibility
 
Well, then you have no bigger website.
Because you miss there the most important factor - speed.
And no, its not cpu.

Sorry, no offense. But look at WF itself.
You can't really speed up a site like this by renting bigger systems.
The optimation possibilities by hardware are saturated at some point. Or better - its scales up - and scaling up in hardware means money. You would have to invest alot money to make this system faster by hardware.

Thats why you get alot by using CDN, cache or other software and architecure related factors.

On a higher access server the most important goal is to reduce the access time - you want minimize the time your server is dealing with stuff and you want a fast response.

Its not the cpu - in a heavy multi threaded enviroment faster CPU scale massivly down because the power is parted so much - you would need buy ALOT more cpu power to compensate it and thats one of the major scaling points.

If your sotware enviroment sucks then it sucks at all.

There is btw an example of this problem we all know.
Google.

If you want do something like google - a search engine:
Would you use mysql or any other rational DB?
The fact is - they would work good on the logical part. But you can't.
You would scale up so fast that every hardware of the world would not
give you a responsive SE.

Thats why you have to use there something like NOSQL.
The name nosql btw is a kind of joke related to the problem i described.

But again - no offense. For the usual high CPC site with normal traffic there is no need for it - there you are fully right.

And yes - you are right about the affiliate monitor. There it is not needed.

But i developed that program not for that.
The original program does some more - i pull domain infos. ALOT domain infos
as service on a linux server - with alot threads using little cpu time and little resources.

Compare nginx with apache. There you get the same idea.
 
Status
Not open for further replies.
Top Bottom