Writing drop catch software

[monaghan]

I'm considering adding DAC access to my Nominet membership, reading around the subject, would I be right in thinking that it's simply a case of reading your database for the day's target(s) and hammering requests into the DAC (within the contractual limits) until it pops put with an unregistered reply? (or someone else beats you to it). Is there anything more scientific than this?

I presume the domain drops at the same time of day as it was registered (subject to Nominet's systems not having a backlog) thus giving a spread of released domains over the day?

I appreciate that the quality of caught names is only as good as the quality of your lists, but that's another subject and I'd not expect anyone to be sharing their lists

 

[grantw]

Domains are dropped randomly, not the time they were registered.

Grant

 

[bb99]

Yes essentially that's all it is (save for what Grant says - they drop randomly during each day).

The key is optimising the script once you've nailed the basic functionality

 

[monaghan]

I would have thought you need to throttle a script down to the limits in the T&C's rather than optimize it

 

[Jordan]

I would have thought you need to throttle a script down to the limits in the T&C's rather than optimize it

If you say so!

 

[monaghan]

1000 queries per rolling 60 seconds is only 16.6 queries per second, not a lot for a modern system on a fast network to cope with I'd have thought?

 

[Jordan]

1000 queries per rolling 60 seconds is only 16.6 queries per second, not a lot for a modern system on a fast network to cope with I'd have thought?

No, it's absolutely nothing for a server/network but if you think that's all there is to it you're in for a shock

 

[monaghan]

Thanks for your kind words.

I'm aware that there's more to it than just stuffing domain names into the DAC queue and getting rich

The domain reg stuff is already written as an interface to my hosting billing software so can be re-used quite simply and databases are my main day job, so pulling a nicely optimised database of names would not be a problem (populating it with good names will be the challenge ). And on a current PC / Server even with a basic scripting language, there should be plenty of time left in that 16th of a second to run the housekeeping that's necessary.

The main aim of the exercise is "I want to do it" rather than "get rich by catching all the best names"

 

[stevebrowne]

Remember that although the MAX you can query at is 1000 / 60s (so 60ms on a rolling basis), names drop and get picked up within about 15-20ms.

If you are checking every 60ms, you can see the discrepency...

 

[monaghan]

Remember that although the MAX you can query at is 1000 / 60s (so 60ms on a rolling basis), names drop and get picked up within about 15-20ms.

If you are checking every 60ms, you can see the discrepency...

I guess this is where the "luck" comes in - assuming the drops are truely random

 

[Jordan]

I guess this is where the "luck" comes in - assuming the drops are truely random

Who's to say it has anything to do with the 'randomness' of drops (a common assumption) - this is where the 'optimsation' comes in.

Good luck

 

[aZooZa]

Monaghan,

There are a number of factors to consider if you're writing your own software. It takes a lot of time to 'hone' the code and understand what you should be doing. Each individual part isn't rocket science, but the sum of the parts can take some time to get right and tie in together.

Shameless plug:

 

[monaghan]

Thanks,

As I said previously, I'm not expecting to spend 5 mins on the keyboard, compile a bit of C code and sit back and catch all the best domains. This is more of an interest and if I can make it work well enough to grab a few half decent domains, then that's a bonus.

As you say, opening a database and pulling a list of domains, opening a socket and bunging the list at it & checking the result and then if necessary sending an email, each in their own right is simple enough to code, tuning each module to work efficiently alongside the other modules will be the "fun" part.

 

[mxm]

ice cream drops from the batcave

Monaghan,

There are a number of factors to consider if you're writing your own software. It takes a lot of time to 'hone' the code and understand what you should be doing. Each individual part isn't rocket science, but the sum of the parts can take some time to get right and tie in together.

Shameless plug:

nice plug aZ... if Monaghan has not bought from you yet... at least other will get the hint!

Thanks,

As I said previously, I'm not expecting to spend 5 mins on the keyboard, compile a bit of C code and sit back and catch all the best domains. This is more of an interest and if I can make it work well enough to grab a few half decent domains, then that's a bonus.

As you say, opening a database and pulling a list of domains, opening a socket and bunging the list at it & checking the result and then if necessary sending an email, each in their own right is simple enough to code, tuning each module to work efficiently alongside the other modules will be the "fun" part.

nice interest... maybe a small amount of revision on the subject in your own project time will result in a lot of saved "fun" coding... think of all the soccer you will miss! and a bit of $ spent on someone who has already done it (god is that a hint too!) and the most important factor of the project *lists*... it is great to see you want to do it and there is no reason why you cannot put the jigsaw together but it is much more *fun* gettting generics that others have missed or forgotten about in the last 20 years...

 

[monaghan]

think of all the soccer you will miss!

Missing the footie is all the more reason to spend the evenings tuning code, can't stand the game