How to find a short domain name | Hacker News Reader

How to find a short domain name

zegup.com

154 pointsdwyer78 comments

A few years back I built Domain Pigeon, a similar tool, that generated web 2.0-style domain names and made them available to visitors. One of my big lessons learned from that experience was that web 2.0 style domain names are NOT good domain names. More often than not people can't remember them and when they can, they misspell it and if they can spell it, it's only because someone had to spell it out when telling it to them.

I recently launched a follow up to Domain Pigeon called Lean Domain Search which I think handles the problem a lot better [1]. Rather than generating random words, it pairs your search phrase with 1,500+ other keywords and instantly shows you which are available. In this way, every domain it generates is a combination of two or more actual words which makes it much easier to remember and explain to people.

By all means explore and see what your options are, but try to pick a name that will be easy for people to remember and easy to spell.

[1] http://www.leandomainsearch.com

I just checked out lean domain search, it actually looks pretty handy. I like how fast it is at generating domain names and checking for availability. One feature that would make it much more useful is the ability to specify multiple words and have it do a combinatorial check on them against your other words.

One other thing, the "LDS" favicon took me by surprise. I had tabbed away for a second, then coming back I saw LDS and thought "now how the hell did I get on Mormon church site?". Only after I clicked back to it did I realize it was your site. I live in Utah, and the Mormon church is very prevalent here, so I don't know if others would have the same thought or reaction. Just thought I'd mention it. Good luck with leandomainsearch

Thanks. The multiple queries per search idea isn't a bad one, but not enough people have asked for it where it makes sense to implement it just yet. And as someone else mentioned, the results are already a bit overwhelming. Having 3,000 or 4,500 or 6,000 results vs the current 1,500 isn't going to help that. I think my time is better spent exploring ways to organize the current results. But if there is demand, I'll meet it.

And yeah, agreed about the Mormon association. They really need to change their name =X

Add me to the list of folk wanting to search for multiple/combinations of words. Two would be sufficient for me. There are a lot of noun + verb combinations (and perhaps noun plus adjective) that make for a good base for a domain name.

Regardless, the tool as it stands is excellent. Thanks.

"web 2.0 style domain names are NOT good domain names. More often than not people can't remember them and when they can, they misspell it and if they can spell it, it's only because someone had to spell it out when telling it to them"

Agree 100%.

I'd like to give my positive experience with leandomainsearch.com.

I used it when coming up with potential domains for a client (that was referred by a VC - a startup ready to launch with a new domain basically.).

Although the client ended up buying a domain (for sale) that I also suggested, I didn't tell them which names I found that were "free" essentially and which names they would have to buy from an owner. Or any pricing. I didn't want to color their opinion by creating a halo around the names available for purchase.

The majority of the ideas I got for available domains were from LDS. (In all fairness I also curated that list based on my knowledge since if you've used LDS you know it presents an overwhelming number of possibilities.)

The clients budget was about $20,000 for a domain purchase. In the end we bought two domains (one the one they wanted and one which was a typo of the one they wanted) for about $3500 total.

In my opinion the LDS names were as good as the ones the client wanted to buy. In some ways I thought they were better (but of course it's not my startup and maybe I was rated them higher because I choose them..).

The idea that I gave them when considering the final list (which took them several weeks to decide on) was to go out on the street and run the names by potential clients and consumers on this particular site and to see which names resonated with them. That allowed them to hone in on a great name at a reasonable price.

In any case even though they didn't go with the LDS name it was very valuable in the process.

Thanks larrys, that's some of the best feedback I've received.

I encourage everyone to do this: find some premium domain names that cost $X,XXX or $XX,XXX and mix them in with results from Lean Domain Search and see if people really prefer the expensive ones. (That would actually make a really good marketing tool/game that I could implement on the site itself.)

I've said it before: as word spreads that there actually are a lot of quality available .com domain names out there, I think the premium domain name industry is going to feel it. As it should.

I'm learning so much from this discussion. I'll throw in my data point (I believe it's relevant):

I ran shell scripts on 10 nodes with fast backbone connections, making DNS queries of all possible 4-letter and 5-letter .com's. Obviously, if the domain doesn't NXDOMAIN, it has been taken.

At the time (before domain squatting was shut down) all legal 4-letter domains were taken.

From the NXDOMAINs, I ran a much slower whois script to see if the domain was registered but missing a DNS entry.

From that final list of unclaimed domains, I was quickly able to spot the one I wanted.

"domain squatting was shut down"

No such thing as what you are saying. You are probably referring to the penalties that are assessed for domain tasting whereby a registrar is penalized if they return to many domains relative to the ratio of registered domains (registrars also pay more money to ICANN depending on how they handle this as well).

Much to much work to do the way you are doing it. You need to get a copy of the zone files there are people with access to that (we have access but we don't sell access others do though for nominal charges).

Thanks for the great insights!

I recently obtained a five-letter domain name (Grovr) in .org/.net/.com, and felt good about doing so on the cheap.

I hit on the that name out of desperation leading to silliness. The idea has to do with directed acyclic graphs of trust relationships, hence trees, hence "grove".

There aren't a lot of real-world analogs, so I'd been struggling to find some short name that explained the concept. I was working with a few others until recently and they had picked names like "layervote" which I never liked much. "Grovr" is obscure, but it makes me happy, and I'll probably make trees the fundamental graphic theme and metaphor.

But now you're telling me that misspellings are actually a big problem? How much of a problem, quantitatively?

Does that outweigh googlability? (On the other hand, if you add one letter, there is a popular children's TV character there which I'll never outrank, so that might be a problem.)

"But now you're telling me that misspellings are actually a big problem? How much of a problem, quantitatively?"

When you verbally tell someone about your site, how do you do it? Do you spell out the domain name? If not, people are probably having trouble finding it because it's not a standard spelling. And looking at the name I'm not even sure how to say it (is it "groover" or "grover"?). If you do spell it out, that should be an even bigger indication that people are going to have trouble because you had to spell it out in order for them to find it.

That's not really quantitative, but I see your point.

Flickr was the prototypical Web 2.0 name, and I always thought the unusual spelling was a plus, since it made it more memorable, but perhaps there were a lot of other potential customers who were lost at step 0.

And I suppose you have similar conclusions about domain hacks, or using dictionary word + obscure TLD?

I suppose I am aware that for the average person, a website address is "www.____.com" or nothing. But I have an irrational love of interesting domain names. Perhaps it's bad to indulge that.

EDIT: The US government has some guidelines for domain name usability here. Usability isn't quite everything, but it's probably worth serious consideration.

http://www.usability.gov/articles/newsletter/pubs/032007news...

Yeah, definitely avoid domain name hacks. It's a huge gamble and even if your product is successful, it probably would have been more successful with a better name.

I thought the 'web 2' domain names came about, as most dictionary words had been taken. Take a dictionary word and make a small alteration - so that it is no longer a dictionary word - and it's likely you have something unique.

Search engine auto-correction must be an issue though!

I'd probably pick obscure word combos, that wouldn't naturally occur together. Good fun trying them out in a search engine too.

Nice site. Can I request TLD choice? I'd use com/net/org/co.uk

No, Lean Domain Search does not support other TLDs for the same reason you shouldn't use domain hacks: it degrades the quality of the domain name and confuses normal people.

It confuses normal US people. I'm pretty sure normal British people are perfectly comfortable with .co.uk, Australians with .com.au, etc.

Ah, ok, so you wouldn't recommend a .co.uk for my UK-focused product then? Interesting. Oh wait, no, the other thing, useless.

www.nametoolkit.com.

We have lots of different algorithms, and you can click on one to get just those results.

Am I right to be suspicious of this service if the best name it could find for their own domain was zegup?

>babof.com is available

Yey !

so do babof, babef,...

> best name

Best (1) currently available (2) 5-letter domain name. I think they did well for these standards.

True, but it's not really that necessary a restriction unless there's some reason that five characters is special. I mean, would you rather have mxurp.com or something meaningful like ycombinator.com?

How do we know that as these searches are executed that the domain names don't become unavailable, because someone parks on them? That historically has been the problem with domain-name searches. I'd like an absolute guarantee that the searching is done in a manner that doesn't alert registrars or their cronies to new domain names to park on before running many searches here.

The application is a text-searching application, so I was a bit startled to see "consonant" misspelled as "constonant" throughout the pull-down menus. Usually I expect programmers who work with text string-matching to be attentive to details of natural language text.

I'm not seeing anything on zegup which allows you to even indicate from the list which domains you might be interested in. (Unlike LDS where you can go and register and click on something). Not that they won't add some way to mark a domain of course or allow you to register.

For speed purposes they should be FIRST running this check on yesterday's zone files from each registry (Verisign for .com/net for example). If that fails (shows in the zone file) no need to go further.

After that check if a name does show as available you do a whois against the appropriate registry NOT a REGISTRAR (because there is no reason to do that).

As step 1 if you don't have access to the zone file you can do a nslookup to see if the domain resolves. If it doesn't you can then check the registry (by the way PIR (.org) rate limits checks unless you are white listed).

Theoretically if a registrar has access to zegup they could front run domains IF the user of zegup indicated which domains they were interested in. Otherwise to many names to be of value in front running the way I see the site operating.

Of course there is nothing to prevent a registrar from putting up a man in the middle site (like zegup) hosted elsewhere and grabbing domains and doing what you are saying. So there is no way to "guarantee" what you are asking other than the reputation of the person running the "zegup" type site.

As an aside zegup.com was registered recently and has domain privacy. So even though whois can be faked privacy info when you are runnning a business is a bad idea.

Domain Name: ZEGUP.COM

Registrant: PrivacyProtect.org Domain Admin (contact@privacyprotect.org) ID#10760, PO Box 16 Note - All Postal Mails Rejected, visit Privacyprotect.org Nobby Beach null,QLD 4218 AU Tel. +45.36946676

Creation Date: 09-Feb-2012 Expiration Date: 09-Feb-2013

That was my first thought too. Any time I search for a domain name these days, my process goes something like:

1. Enter it into a web browser, see if it's in use or parked

2. If not, run whois on it at the commandline.

3. If not, decide whether I want to buy it. If so, do so right away, otherwise assume it will be lost in a day or so.

And I'm not even 100% sure I can trust the browser search, depending on what the DNS server I happen to be on does with failed lookups.

I personally don't feel that this is a reason for concern (very very small chance but yes it is possible). But there is a faster way anyway, cutting out web server not responding etc.

So in any case if you wanted to lessen that chance you could query any public nameserver, assuming you can trust that nameserver.

In a command window, using nslookup as an example, the process would be (using google public dns server but you can use anything that's the point):

nslookup

> server 8.8.8.8

Default server: 8.8.8.8

Address: 8.8.8.8#53

> skymarshal.com

Server: 8.8.8.8

Address: 8.8.8.8#53

Non-authoritative answer:

Name: skymarshal.com

Address: 208.91.197.26 (comment: name taken!)

> skymarshalxyz.com

Server: 8.8.8.8

Address: 8.8.8.8#53

server can't find skymarshalxyz.com: NXDOMAIN

> www.skymarkshalxyz.com (I made this up)

Server: 8.8.8.8

Address: 8.8.8.8#53

server can't find www.skymarkshalxyz.com: NXDOMAIN

> exit

"can't find" - name might be available - we have determined it's not resolving. Could be a dns issue of course but it's a good first step.

Doing with a browser can give you MORE of a false positive. DNS could respond with an address but the site could be down. Of course getting no response to DNS doesn't ALWAYS mean the domain is available but it's generally fairly accurate. At that point you would do a whois to confirm.

You can write a shell script of course to automate this from the command line.

You can safely use http://instantdomainsearch.com. It was written by a YC alum. Lots of companies we've funded have found names there.

That appears to be using nslookup (edit-"esq") to see if the domain resolves as a result of a test that I just ran. (Tried a name not in the zone file that IS taken (4L .com) - edit: it said it was available. Then tried a name that wasn't in the zone file and it took some time to determine if it was available. If it was comparing to the zone file there would be essentially no delay it would be a simple search. If I wanted to prove this I could register a name NOW, do the dns, and then do a query but I don't have time to do that. The zone file is only available once per day for download.

Also, it doesn't work for anything other than .com.

And if you hit "search" it sends you to godaddy.

All in all I think it makes more sense to simply find a dns server that you can trust OR get a hold of the zone file and query that (safest way actually). My personal feeling though is that this is all a non-issue.

May I suggest they change the default registrar from GoDaddy to something else?

I felt a bit limited by this, so I've just knocked up a similar tool with far more flexibility. Still actively developing it, but it's far quicker already.

http://dev.mindfuzz.net/domain/

Edit: Grammar.

Much better. I like this. Reminds me of when I did the same thing years ago with regexs to grab 3/4 letter .com (really there were so many available back then you had to whittle the list down).

Thanks! I'm looking at adding more tld's and words/synonyms.

This returns lots of false positives. Certain patterns, LLLL, result in very small result sets (5 items) but when re-searching the pattern, a different set is returned.

Yes, because I use more of a Monte Carlo approach than a systematic approach. zegup.com is always alphabetic, but since the results are so slow it could take a long time to get to the end of the alphabet.

+1 for randomizing.

do you produce the string with the monte carlo choosing within a specific subset? I'm not sure why it wouldn't produce more than 5 combinations and then 5 different combinations...

also, if you have-time/like, you can probably implement a quick hash on the fly for the domains already returned to avoid listing the same domain twice in a search with few alternatives, it should be quick.

but cool tool!

Hash is a great idea. Thanks! I'll implement the hash now.

I build an array of arrays where each subset array has all the possibilities for that letter. So if the query was N then a[0] = [0,..,9]. Then I just pick a random element from each subset array and stitch them all together in order.

Great tool. Returned some false positives for me. Also, searching for a specific domain in all lower case will "find" that domain 50 times.

Ah yes I completely forgot to stop duplicates. I'll just add that now.

Checking if a domain name is registered can be a trick task. The only reliable method is to run a whois; but most hosts only search a fraction of the whois records (perhaps just .com) and also limit your queries. You need to pay money to make mass requests.

I have opted to check for any kind of DNS record. That means registered but unused domains will still return available, but the trade off is that it's very quick, pretty accurate most of the time and most importantly free. :)

i like this.

the page is pretty, results are fast, and the letter based slots specification is nice.

but i think the results are pretty low quality (gibberish).

Thanks for your comments. :) I agree that the quality of the results are not of the best quality, but I think it can used to source some interesting information.

Warner Brothers appear to have registered batman[1-9].com, heh.

You should be focusing on the quality of the domain and not the length. This point is clearly illustrated when you use this service, as almost every domain generated is awful. While it is true that a domain can be too long, you will certainly find better domain names by looking around 8-14 characters instead of 4-7.

Takes so long to generate first 50 names even when searching with just the defaults... looks like they don't really cache anything..

Cache random results?

No such thing as a random result when you have static selection boxes defining specific input. The results are obviously restrained by what letter combinations you pick (ie finite).

You could make an argument that caching all the possible combinations would be unreasonable for what's likely a low cost setup. However, I'd argue in favor of caching with both a popular and a simple reduction scheme.

Could speed it up radically by just caching on the first letter / number.

It should be possible to randomize. I'm only getting domain names starting with ba.

I got this error a little later, when the script was already running for a while:

Error Connecting To Server: whois.crsnic.net

Maybe you should not use "whois" but just a simple "host" or "ping" to resolve whether or not a domain is free. When in doubt, you can still do an "whois", but for a heuristic approach, it would be waaay faster.

Faster, but inaccurate. There are plenty of domains that are registered, but the DNS doesn't resolve. That doesn't mean they are available to register!

When I wrote a domain searcher script what I did was filter first by names that return are resolvable, and then whois the remainder. Reduces load on the dependent whois server whose terms of service usually say "don't run scripts against me"

That's exactly the approach I tried to suggest: Find as many correct replies first, because there certainly are hosts which names can be resolved and try the whois-approach on everything else.

do you think that would return too many false-positives? many of these domain search tools use an api for whois or for searching.

It's probably perfectly honest but this almost feels like a turk exercise to get other people to find short names for you.

Bah, it doesn't even randomize the access. I put consTonant consTonant and of the million possibilities it started checking all the bb___ ones... not so useful if they show just 50 and the only thing they allow you to do is to keep going with the next 50...

I wonder who designed the algorithm...

Related: http://messymatters.com/nominology (discusses desiderata for names: evocativity, brevity, greppability, googlability, pronounceability, spellability, verbability)

http://www.kevinohashi.com/17/04/2011/other-24000-available-... 24000 5 letter 'brandables'

Sorry if this is a dumb question, but what's up with the letter Q?

Maybe it's like scrabble where you'd always want a u afterward.

It's a rather odd constraint given the other options.

this thing outputs crap. every single name/domain name generator i've seen has been underwhelming, including this one.

do this: * use an emotional connotation dictionary to classify a bunch of poems. * for each emotion category train a markov chain. * allow users to select a couple of emotions, then combine those emotion's markov chains and generate text.

I have an iPhone program called "Inspiro" which generates phrases, story ideas, etc. Great for improv or a random laugh. Got my blog domain idea "syncing dreams" from it, and I'm happy with that. No, it isn't checking for domain availability of course, but the combinations are unusual enough that you'll probably find the domain available. A few Inspiro results:

charmless koala

whispering playgrounds

day-glow demons

juicy naked mole rats

forging ice

You can generate different kinds of nonsense, like:

ruthlessly sneezing on perfect phrases

cheerfully grabbing handfuls of fascinated buildings repeating naughty garden hoses

So, fun for strange ideas that make your brain get creative. My favorite tool for generating actual domain names though is easily http://impossibility.org/

that sounds pretty good.

this is kind of a tangent but kind of related: you can use a reverse dictionary (http://www.onelook.com/reverse-dictionary.shtml) as a sort of "multi word thesaurus" when trying to condense a couple of ideas down into one word

This is beautiful, but it's such a shame that US law enforcement have been stealing .com domains from people for things that their users have done. We've had a bunch of Hacker News threads and I don't want to resurrect old beasts of that sort, but some option for non dot-com names would be wonderful. (Especially because .com reminds me of 1999-2001 at this point and it seems like we designers should perhaps be treating it as a ten-year-old fashion.)

I couldn't disagree more with that we should be treating .com as passe. While .me or whatever is "cool", the majority of the world still relates to .com's the most. I'm in Taiwan right now and the people here probably have no idea what a .me or a .it is. They are using .com.tw, not .me.tw (and yes, of course I understand that .me is from Montenegro and this would never be possible). And the "reminds me of 1999-2001" just sounds silly...it goes without saying that there have been thousands of successful .com's since 2001, like the one you are looking at right now.

Actually .me.tw would be possible, just as there are .gov.tw etc. so .com in the .com.tw is not a top-level domain (.tw is), it is a 2nd level domain in this case. so there could be xyz.tw just like .co was originally exclusive to government of colombia and you could only get 2nd level domains, ex domain.com.co etc.

thanks for clarifying, I edited-in that line in () mainly to qualify that I was sort of joking about the .me.tw. I now know it wasn't a joke.

I agree. A quick glance at the top 5,000 domains in terms of traffic more than demonstrates the sheer dominance of the .com brand for domains. It's beyond overwhelming.

There are very few relevant .me or .io (etc) domains. That will continue to be the case for years to come.

It's not Twitter.me or TwitPic.me or yfrog.me, it's not Tumblr.me, it's not Pinterest.me, it's not Groupon.me, it's not homeaway.me, it's not yelp.me. All the major success stories online will continue to be branded primarily through .com - and those that start out as .me or .io will move to .com as they can afford to purchase the prime domain.

This, I was gonna say the same. Who wants a .com? Not me

On some domains I get a invalid domain message:

bbaaw.com is an invalid domain name

Any ideas on why this is appearing or what it means?

looks like a bug

This coupled with the importXML feature of Google Docs is very useful.

Does it mean that all 4-letter-.com domain names are taken?

When they all were taken mainly by domainers they were going crazy that they'd be the new LLL.com's where the prices went from $xxx to $xx,xxx in a very short period of time.

However, what has happened with LLLL.com's is that they went very quickly to $xxx but have since dropped considerably so you can acquire LLLL.com's (4 letter.com's) from $15 upwards although if you want a really nice LLLL.com such as Made, Ball etc then those prices are different regardless of it being an LLLL or not.

Just have a look on Domainer forums such as Namepros, DNF, DigitalPoint etc and you can find some really cheap.

Yep, all 4 letter domains were registered back in 2007 (http://www.dotsauce.com/2007/11/02/four-letter-domains-all-r...).

Of course domains expire or are sold between owners so getting a 4 letter domain is still possible.

And you can get them for cheap. Got one .com as recently as last year, with the .org as a freebie. I'm now considering selling my 3 letters/numbers domain name - .net and .org.

How much was the .com? I've seen offered prices of $5K+ for fairly generic 4 letter domains...lot's of speculators buying and holdnig I guess...

Speculation- yes. I paid just below 100, and I considered that a tad expansive.

It's not a word but it is easy to say and to remember, and I needed one for a project, so I got it thinking I could reuse it later.

That's not how you spell consonant.

We have many short domains: if you want to take on holly wood- you could use Free.TV or if you want a communication startup, try XG.org, Free-fi.com, wusb.com, or wide-fi.com. News site? Syndi.com, bloog.com. Hyperlocal? 4ny.com or 4dc.com. Keyword? Liposculpture.com non-profit? Manna.org My favorite for Big Data is NoSQL.com

While this post is a plug for my open source project OpenDomain - please note that we are not for profit. We have given domain such as Drupal.com, OsCon.com, OpenAjax.org, and Ecmascript.org all for FREE

EDIT: spelling errors edit 2: please do not downvote - as I said before we are not for profit and we support open source. I would love for anyone on hacker news to use any of our domains

Reply on news.ycombinator.com