Membership is FREE, giving all registered users unlimited access to every Acorn Domains feature, resource, and tool! Optional membership upgrades unlock exclusive benefits like profile signatures with links, banner placements, appearances in the weekly newsletter, and much more - customized to your membership level!

Nominet say can have as many TAGs as like.

Status
Not open for further replies.
Jay Daley said:
If a developer chooses to use SMTP command X instead of SMTP command Y because they think it shaves a few milliseconds off then what business do we have telling everyone else that?

and

Jay Daley said:
There are some things mentioned in this thread that I only learnt about when I read them here - such as using a single character for address data

With respect (because I do respect Nominet):

There are many things I think you don't know. Not dishonesty, just naivety.

and if Nominet supplied the client software you wouldn't have that anticompetitive dilemma.

-aqls-
 
Last edited:
On the subject of prenoms, please can someone at nominet towers press the magic button.
 
mreeves said:
On the subject of prenoms, please can someone at nominet towers press the magic button.

Agreed, well overdue! I've got my de-randomizer ready to go ;)

Grant
 
De-randomizers are sooooo old school. That is where you are going wrong Grant. I suggest you upgrade to a quantum powered crystal ball ;)
 
I think the fairest way would be for Nominet to provide the software/script for making the requests so that only their software/script may be used, that way the playing field would be levelled. I dont know if this is technically possible though, is it ?. Nominet could then provide this and members then MUST only use their provided script/software ?? Bit like "shutting the stable door...." but hey better than nought.

DG
 
mreeves said:
suggest you upgrade to a quantum powered crystal ball ;)

That was next on my list but I hear tagnames have employed Mystic Meg - may as well give up now :)

domaingenius said:
I think the fairest way would be for Nominet to provide the software/script for making the requests so that only their software/script may be used, that way the playing field would be levelled. I dont know if this is technically possible though, is it ?. Nominet could then provide this and members then MUST only use their provided script/software ?? Bit like "shutting the stable door...." but hey better than nought.

DG

You want everyone whos spent years and maybe thousands of £ on their own systems to stop using them just because your system doesnt catch the domains you want???

Grant
 
vizzy said:
I believe what Nominet should do at this stage, is to implement proper truly random algorithm for prenoms drop cycle.

I can't understand why they have to drop them in a favourable way for some people who, due to one or other reason, know the pattern.

As I've said before: The day on which prenoms are dropped is chosen by a senior manager based on whether or not all avenues are exhausted in trying to find the current registrant. The names then enter the standard random queue for dropping.

This personal decision prevents someone making a false claim about a prenom and then withdrawing it all of a sudden in order to try to manipulate the time it drops. Every effort is made to make this unpredictable.

So only one person know what day, which has no pattern to it, and then nobody knows what time on that day.
 
grantw said:
That was next on my list but I hear tagnames have employed Mystic Meg - may as well give up now :)



You want everyone whos spent years and maybe thousands of £ on their own systems to stop using them just because your system doesnt catch the domains you want???

Grant

YES ,got it in one. Why should those with the money and know how to do so have any benefits over someone without when the benefits to be gained are a public resource,in short they should'nt. Or are you saying that just because YOU have a good system no one else should be allowed in unless they have the money to develop a similar system ?. There is capitalism and free market ,and there is UNFAIR capitalism and free market and what I point to is that this is not some private deal between yourself and a private company, this is a deal between yourself and an organisation that is SUPPOSED (I thought ??) to be there to operate a fair system for ALL the public. IS it too much to ask that Organisation,Nominet, to bring in a system that means EVERYONE is on a level playing field ?.I think not.Auction the dropped names so that everyone,including people that want to actually use the names can have a chance. Was Nominet brought into existence to hand 75% of the domain names on a plate to a handful of speculators ?.I think Bernd-Lee would turn in his grave !!

DG
 
domaingenius said:
YES ,got it in one. Why should those with the money and know how to do so have any benefits over someone without when the benefits to be gained are a public resource,in short they should'nt. Or are you saying that just because YOU have a good system no one else should be allowed in unless they have the money to develop a similar system ?. There is capitalism and free market ,and there is UNFAIR capitalism and free market and what I point to is that this is not some private deal between yourself and a private company, this is a deal between yourself and an organisation that is SUPPOSED (I thought ??) to be there to operate a fair system for ALL the public. IS it too much to ask that Organisation,Nominet, to bring in a system that means EVERYONE is on a level playing field ?.I think not.Auction the dropped names so that everyone,including people that want to actually use the names can have a chance. Was Nominet brought into existence to hand 75% of the domain names on a plate to a handful of speculators ?.I think Bernd-Lee would turn in his grave !!

DG

When I started in this I had no money and no idea how the system worked - I took the time to learn and develop my own system as did everyone else on here - I didnt start moaning 6 months in because I wasnt getting any domains, I worked harder to understand why I wasnt getting them and amended my code. Domains are not being handed on a plate to anyone, they are being deservedly taken by the people who could be bothered to spend years of their life working for them.

grant
 
Right folks, we've just done another thorough investigation into some names that it was suggested were unfairly caught, defying the laws of statistics etc etc. Once again the answer is very, very clear. The names were picked up by an email received after the names dropped, following a DAC query, no specs were sent before and the person who got them beat the competition by a very short margin.

So to be as clear as I possibly can, there is no funny business, no skulduggery and no cheating. Some people just have systems that are consistently faster than others.

As far as I can tell none of the speed improvements relate to undocumented features in our systems. They are about how programs work, how networks work and how SMTP works, none of which is my business to start explaining. There are plenty of good books and references available.

Also, whilst I have a lot of sympathy with the view that we should auction off names, that is not our policy. It has been discussed several times before and always rejected. If you would like to see it as policy then you need to approach the Policy Advisory Board (which is independent of Nominet) and try to get them to recommend it.

Finally, in the future we may well write some client software that shows how to connect to our systems (such as when EPP is available). Some of this might even be good enought to be put into production with some work. However we will never write something that can be used by dropcatchers. This is like an arms race. When one side thinks of a new technique then the other side works out how to leapfrog them. Just keeping up require huge effort.
 
365 said:
Jay,

If one SMTP command has an advantage over another and you know that, then in the interest of having a fair system, document it is exactly what you should do and if you don't then by very definition it is an undocumented feature.

How you send the email is up to you and has nothing to do with Nominet.

Therefore, the 'best way' to send the email to them is not a feature, whether documented or not. :rolleyes:
 
OK, thanks Jay for that - your time and attention on this board is appreciated.

(and I'm just off to get my first and surnames changed to '-'.)

-aqls-
 
Last edited:
Jay Daley said:
Right folks, we've just done another thorough investigation into some names that it was suggested were unfairly caught, defying the laws of statistics etc etc. Once again the answer is very, very clear. The names were picked up by an email received after the names dropped, following a DAC query, no specs were sent before and the person who got them beat the competition by a very short margin.

So to be as clear as I possibly can, there is no funny business, no skulduggery and no cheating. Some people just have systems that are consistently faster than others.

This just doesn't ring true.

Let's say I'm looking at three names today and so are a lot of other people. Nominet say that the names will drop at random. So we are all testing each name, once every 1200mS.
So there's a big element of chance as to who sees it drop first and provided that your system reponds reasonable quickly (say under 100mS) the chances are that you will get your request in first.

Now, if you knew roughly when the names would drop and in which order, then you could run your system in quick bursts and just go for one domain. Then a quick system would help.

Any sign of that happening Mr Daley?
 
FC Domains said:
This just doesn't ring true.

Let's say I'm looking at three names today and so are a lot of other people. Nominet say that the names will drop at random. So we are all testing each name, once every 1200mS.
So there's a big element of chance as to who sees it drop first and provided that your system reponds reasonable quickly (say under 100mS) the chances are that you will get your request in first.

Now, if you knew roughly when the names would drop and in which order, then you could run your system in quick bursts and just go for one domain. Then a quick system would help.

Any sign of that happening Mr Daley?

good question :)
 
Say for example there are 1200 tags and for example they all make a request in order at 1 ms intervals (ok, its just an example).

When the domain drops between the time your script hears about it and sends a reply which is:
Time taken for Nominet to process email to you +
Time taken for email to get to you +
Your script time (100ms) +
Time taken for your email to get back to Nominet and add a timestamp.

Which for sake of argument let's say is 200ms.

Upto 200 other tags will have heard about the free domain (ok quite alot less even in this example).

If their script is running at 10ms compared to your 100ms and their email is optimised to be much quicker than yours then they will have 80ms plus email optimisation time in which to get their request in before yours.

So even if they hear about it 79 ms after you did they will still beat you.

Corrections please?

-aqls-
 
aqls said:
Corrections please?

-aqls-

No corrections, but that is an extreme example.

However even with fairly long delays the fastest only stands a slightly better chance and the results should be more distributed.

A lot of people are reporting 30mS responce. So somebody with 0mS should only be less than 10% better off.
 
Cheating V Good systems

Just to poke my nose into the trough...

It strikes me that those who think there's foul play won't rest until they're actually told precisely how the top catchers are able to do so well. This isn't going to happen because, for most, their code, their optimisations and their system are all commercially sensitive.

I have found myself wanting to comment on a lot of the posts in this thread but have decided against it on the basis that it would give away optimisations that required a lot of very hard work to research and implement.

Nominet have many policies relating to "abuse", whether that be abuse of WHOIS data or abuse of DAC or Automaton limits. Nominet also have a very strict "don't f*ck with us" policy [1] when it comes to this stuff. There are recent court cases to back this up. Not on the DAC, but on WHOIS etc.

Nominet closely monitor the usage of the DAC etc and they do investigate complaints when they're made.

If you're a Nominet member and you don't like "the system" as it stands then the best course of action is to take it up with one of the elected PAB Members as they are best placed to review the relevant policies and, ultimately, push through any change.

I find it annoying when names I go for get caught by someone else. Really annoying. But I trust Nominet to police the system and instead of busting everyone's balls about it I channel my energy into improving my chances of catching the next time.


[1] This may not be the correct legal terminology.
 
Last edited by a moderator:
Perhaps Jay you could advise me roughly what percentage of the .co.uk names are held by speculators and what percentage by "real" end users ?. From my research I would guess about 75% by speculators and 25% by real end users ?.

Any ideas ?

Inshallah
 
I may be barking up the wrong tree here, but if people aren't firing off ANY speculative registrations but are waiting until the DAC reports the domain as available, they don't have to limit themselves to one registration email? For instance, they could fire off a batch of 100 and hope that one of them gets there quicker than anyone else's do... that still leaves them with 900 in-hand for the 24 hours in question.
 
Status
Not open for further replies.

The Rule #1

Do not insult any other member. Be polite and do business. Thank you!

Premium Members

New Threads

Our Mods' Businesses

*the exceptional businesses of our esteemed moderators
General chit-chat
Help Users
  • No one is chatting at the moment.
      There are no messages in the current room.
      Top Bottom