Talk:MaraDNS/Archive 1

Just to show people I support the no self-promotion edict, I, the author of MaraDNS, will not expand this article. Samboy 05:08, 5 Jun 2005 (UTC)
 * Actually, I've changed my mind. Once 1.1 hits beta (projected date: Early November), I'll expand this article. Samboy 07:31, 6 October 2005 (UTC)
 * As long as you observe NPOV guidelines, you should be OK. In any case, there are a few people that watch this article, and may revert or edit any contributions that are deemed to be biased. I don't get the impression that you'd do that, though. Mindmatrix 20:20, 6 October 2005 (UTC)
 * Done. Maybe this expanded article will get more people to test the 1.1/1.2 branch, which will result in a better 1.2 release.  Also, by commiting to releasing 1.2 in December here, I am more likely to actually release MaraDNS 1.2 this year (The 1.1 branch has had almost as many delays as a Microsoft release).  :)
 * Talking to myself, but I think the main NPOV problem here is that this article is longer than the BIND article (the djbdns article has the same problem). Samboy 08:44, 9 October 2005 (UTC)
 * That's not a NPOV problem, that simply represents a lack of contributors to the BIND article. So long as the contents of this article don't make any unwarranted, and unverifiable, claims about MaraDNS, there's absolutely no problem. BTW:, I'd be careful about saying things like this expanded article will get more people to test the 1.1/1.2 branch, since that could be interpreted to mean you're using Wikipedia to advertise. As it stands right now, the article doesn't read like an ad, so there's nothing to worry about. Just thought I'd mention it, just in case. Anyway, I also thought I'd mention that I'll probably take a stab at copyediting the article, once you've added what you need to. Feel free to leave a message on my talk page, or just leave a message here. Mindmatrix 02:16, 14 October 2005 (UTC)

Some minor corrections:


 * "1.2 will be fully compatible with 1.0 data files"

Actually, if there is a misspelled parameter in a 1.0 data file, this parameter is ignored in 1.0 and causes a fatal error (with the name of tha parameter causing the problem) in 1.2


 * "MaraDNS can resolve any site that other DNS servers can resolve"

Actually, in the very rare case where:

MaraDNS will not be able to resolve the host name in question. Samboy 00:23, 14 October 2005 (UTC)
 * You have a glueless NS referral
 * And this referral which resolves to a list of IP records
 * And the first IP in that list refers to a dead host


 * Clearly, I don't know my DNS terms nearly as well as you do; what does glueless NS referral mean? I'm mostly curious about the glueless part...
 * BTW: did you post the above to discuss how to change those sentences, or were you identifying problems and asking for someone else to re-write the relevant portions? Mindmatrix 02:16, 14 October 2005 (UTC)


 * Yeah, well, writing a DNS server has made me learn more about DNS than I ever wanted to know. About a "glueless NS referral", I feel a Wiki article in the making here:

When a recursive DNS server requests for the name of a domain, they send a request roughly like this:

"I want the IP for www.example.com."

Assuming an empty cache, they will send the above question to a root DNS server.

The root DNS server will tell the recursive cache something along the lines of: "I don't have the answer, but the .com servers are named 'ns1.root-servers-for.com.' and 'ns2.root-servers-for.com'. 'ns1.root-servers-for.com.' has the IP 10.1.2.3.  'ns2.root-servers-for.com.' has the IP 192.168.55.67"

The recursive cache will then ask 10.1.2.3 or 192.168.55.67 "I want the IP for www.example.com"

The .com DNS server will answer in one of the following forms:

"I don't have the answer, but the example.com. servers are named 'a.ns.example.com.' and 'b.ns.example.com.'. 'a.ns.example.com.' has the IP 172.18.4.5.  'b.ns.example.com' has the IP 10.55.32.44."

This is called a glued NS referral; the glue is the "'a.ns.example.com.' has the IP 172.18.4.5. 'b.ns.example.com' has the IP 10.55.32.44." part of the answer.

However, the .com DNS server may give us a glueless NS referral:

"I don't have the answer, but the example.com. servers are named 'a.ns.example.net.' and 'b.ns.example.net.'"

In this referral, we are only given the names, and not the IPs of the name servers; the recursive nameserver now has to go back to the roots to get the IPs for these nameservers before continuing to resolve the name in question.

As for the rest, yes, please rewrite to make more accurate; I glossed over details to make the description simpler (an encyclopeida article for a fairly obscure open-source program should be short and concise). Samboy 05:58, 14 October 2005 (UTC)


 * Thanks for the explanation. I knew about the dynamics of the process, but I had never heard the term glueless to refer to it. I guess that's my new fact for the day :-)
 * In terms of the article itself, I agree that being concise is important, but that doesn't mean the article needs to be short. The fact that the project is obscure doesn't mean a comprehensive article cannot be written about it. Wikipedia has lengthy articles about obscure places in the world, for example. I don't think this article needs to be lengthy, either, but there's no reason to keep the article short. Here's what I think can be done:
 * create a short History section
 * perhaps describe why it was created, when other similar tools existed
 * provide a list of primary features
 * provide a comparison of major features between MaraDNS and BIND, PowerDNS, djbdns etc.
 * That's just off the top of my head. I'm sure there's other info that can be added. Does that seem reasonable? Mindmatrix 14:57, 14 October 2005 (UTC)


 * I think the place to start is to make sure the BIND article is nicely fleshed out. I don't think it is particularily balanced that the djbdns article and the MaraDNS articles are larger than the BIND article; about 85% of the domains are served with BIND; about 5-10% of the domains are served with djbdns, and the rest of the domains out there are served with other DNS servers.  The size of articles should correspond somewhat with the popularity of a given DNS server; and not be adversely affected by which DNS implementor happens to be a Wikipedia editor.  Samboy 22:29, 16 October 2005 (UTC)

I fixed the DNS link which previously pointed to a disambiguation page. But should it be the full name or the abbrevation which links? Kasperd 14:50, 22 February 2006 (UTC)

Wikipedia is very useful
Hey guys, I just want to thank Wikipedia for being here while the MaraDNS.org web page is down. Since this is the second link that comes up when people Google for MaraDNS, it's very useful to be able to edit the page to point people to the MaraDNS blog while we work on getting the web page up again. Samboy 23:14, 22 May 2007 (UTC)

MaraDNS and COI issues
There have some concerns about my edits to MaraDNS because of WP:COI. That in mind, I will restrict my edits to updating the "current version" whenever I release a new stable version of MaraDNS. Samboy 22:36, 20 September 2007 (UTC)


 * Changed my mind briefly, since another editor requested that this article be given references. There is a definite WP:SPS problem here, but I think the article is a little better.  Samboy 21:15, 2 October 2007 (UTC)

WP:NOT
As a heads-up: I can't find any documentation of major enterprises or service providers using this software, or any evidence of widespread deployment in smaller enterprises. The first page of Google results point either to the author's pageso or this article. I don't want to flag the article as NN without asking where I can find reliable secondary sourcing to establish notability.

Per WP:NOT, the primary criteria for notability is reliable independent secondary sourcing. Entries in "glossaries" are not reliable sources.

--- tqbf 04:14, 21 September 2007 (UTC)

Moving beyond WP:SPS
In the interest of moving beyond WP:SPS, here are some links mentioning or discussing MaraDNS: http://www.tisc-insight.com/newsletters/42.html http://www.irccrew.org/~cras/security/software.html http://www.gluon.hu/dns_benchmark

I think the article can be improved by using links like these. Samboy 21:29, 2 October 2007 (UTC)


 * All in favor of improving the Wikipedia, but, in the order you presented these links: (1) is not independent: you wrote it, it's virtually a copy of material from the MaraDNS web site, and so has nothing to contribute to the article beyond your own website; (2) makes practically no mention of MaraDNS, and is on a page the author says, "isn't really being updated. Maybe I should just remove it completely"; (3) makes no textual mention of MaraDNS, but rather lists among a table of other name servers without comment.


 * And hence my WP:NOT concern. --- tqbf  22:06, 2 October 2007 (UTC)


 * WP:SPS does not matter when something is published somewhere else. In other words, the exact same words that are on my webpage no longer have WP:SPS problems when published somewhere else.  Which is exactly what happened.  Between this article and BLACKLISTEDURL/article/269641/maradns_an_easy_to_use_open_source.html, we have enough material for a small, but decent Wikipedia page.  Add a link to the current version from the MaraDNS page and a mention of the security advisory, and we have a well-fleshed out article.  Samboy 21:23, 3 October 2007 (UTC)

Some other brief mentions of MaraDNS:


 * http://lwn.net/Articles/19954/
 * http://linuxmafia.com/faq/Network_Other/dns-servers.html

Samboy 21:42, 3 October 2007 (UTC)


 * You're citing SPS. What about WP:SELFPUB? It's a dubious source, it's written by you about your own project, and the article is largely based on it.


 * Regarding your other sources: neither qualifies as an independent reliable significant secondary source, because both simply list MaraDNS among many other open source DNS servers, without additional comment. Nobody is disputing that MaraDNS exists, only whether it is notable.


 * Here's my question: after valiant mutual effort (and thank you), the best things we have to source this article to are:


 * Self-published or autobiographical content in not-notable venues


 * A blog post


 * A mailing list post


 * A passing reference in a Czech? benchmarking page


 * An abandoned web site


 * Have we established notability? I'm just asking; I'm not AfD'ing. On the other hand, if this is enough to establish notability for your project on Wikipedia, there's a lot of other "notable" projects that qualify that we wouldn't be comfortable with.
 * --- tqbf 22:07, 3 October 2007 (UTC)


 * OK, again, the two most notable articles are this one (Yes, I wrote it, but I didn't publish it) and BLACKLISTEDURL/article/269641/maradns_an_easy_to_use_open_source.html. Seriously: TISC insight was not a vanity press; they even paid me to write the MaraDNS article.  Again, I think between two small press articles, we have enough to make a reasonable Wikipedia entry.  As I mentioned to you in private email, the bar for Wiki notability is farily low, especially for free software projects (Yes, Wiki does have an anti-commerical bias for good or for bad); I don't think the Wiki would benefit from having dozens of articles about free software projects removed because they were not "Notable enough".  Quite bluntly, if it's on Sourceforge, has an open-source license, and is usable software, I think the project deserves a Wikipedia article.
 * Why is the bar lower for open source projects? Because they benefit everyone.  I don't get paid for MaraDNS, but people benefit.  It's a great little DNS server for embedded things like routers; I have users happily using it without problem on gateway routers.  It's also a good system for hosting domains; people tell me in private email they host thousands of domains with MaraDNS successfully.  Is MaraDNS the most successful open source project?  No, not by a long shot.  Is MaraDNS notable enough to have a small Wiki article?  Yes.
 * Anyway, we're going around in circles at this point. I think I will work on making MaraDNS a better DNS server so there are even more articles in the press about it. Samboy 23:20, 3 October 2007 (UTC)
 * That sounds like a great plan. Meanwhile: do any shipping gateway router products/projects actually use MaraDNS? That'd be a good thing to get down in this article. --- tqbf  00:00, 4 October 2007 (UTC)
 * People used it on StarOS for a while, but it didn't work out for them. MaraDNS works fine on a router for a handful of clients, but not 250 clients. reference.  And, as it turns out, dnscache didn't work for them either (reference: same thread); only BIND did the job.  And, yes, I have plans on improving the recursive performance. Samboy 14:54, 4 October 2007 (UTC)
 * Talking to myself here, but it looks like they may still be using MaraDNS for their "5 people in an office" router products. The problem, tqbf, is this: The only time I hear about people using my products professionally or in big ISPs is when they send me support email asking for help, or when I look for it myself.  People generally don't install MaraDNS, put it on their system, have it work, then send me an email about it.  Heck, people generally don't install MaraDNS, have it not work, then send me email about it either.  I only found out about the star-os usage from my own Googling.  There are probably other routers using MaraDNS without problem; no one has talked to me about using them.  This is why references are hard to come by, and, yes, this is why the bar for open source projects being on the Wikipedia is pretty low.  Samboy 15:36, 4 October 2007 (UTC)


 * uh...you're saying that dnscache can't handle a mere 250 clients doing lookups? balderdash. i've run dnscache in production both as an ISP and as a business email provider, where it had to handle doing lookups against four rbl's per inbound connection - at a rate of a several hundred thousand inbound connections per day - as well as secondary rbl lookups by spamassassin on a proxy which did a finer 'sift' of the info for scoring after a given message had been accepted. At any given time, my MX's had a few hundred inbound connection attempts going on (most of the zombies attempting to send junk). dnscache would pull maybe 3 or 4 percent of the cpu load at any given time - and this on 400mhz ultrasparc cpus. 250 clients? please. that's child's play. reading the thread you referenced, most of those discussing the matter are, forgive me, utter n00bs. hand n00bs software that is more complex than 'yum install xyz', and yeah, they'll have "problems" with it. oh - i'm just talking to myself too, since this is off topic. Anastrophe 16:42, 4 October 2007 (UTC)


 * Let's me be blunt here: You're not able to be objective about DJB software. I mean, people on the thread noted problems with DJB and you need to personally insult them.  Really: Why do you have such a need to argue with people who have problems with DJB's software.  Why do you get so upset over legitimate criticism of DJB's software.  Let me be blunt here: This attitude you have towards DJB's software means you have just as much a WP:COI issue editing DJB-editing software as I do.  So I humbly ask that you no longer edit any DJB-related articles. Samboy 22:07, 4 October 2007 (UTC)


 * huh? you're kidding, right? i hope you're kidding. because that's an astonishingly ill-informed, ill-tempered, bad-faith, illogical claim you're making, if not. i'm going to give you the benefit of the doubt that you're joking/being sardonic. good one! now, can we discuss the matter seriously? Anastrophe 22:15, 4 October 2007 (UTC)
 * in the interest of fairness, i should 'refactor' my comments about the referenced thread. i didn't represent them carefully. most of those in that discussion were quite simply not discussing either maradns or djbdns. most of those in the thread struck me as n00bz - however, most of those in the thread were not commenting about resolver software. those who did comment about resolver software didn't seem to be n00bz. one person was quite the opposite of how mr samboy (inaccurately) characterizes me - he was in fact suffering DJBDS - DJB Derangement Syndrome ( see BDS for reference). Referencing a thread with rabid anti-djb-software commentary was probably ill-advised at best, i believe. Anastrophe 08:01, 5 October 2007 (UTC)


 * The precedent to this discussion is that you made disparaging and ultimately unsupportable comments about a much more popular competitor to MaraDNS, while at the same time editing a page about your own software to make positive and ultimately unsupportable comments about it. It really doesn't matter what you ask. Improve the Wikipedia and your edits will stand. Scribble all over your better-known competitors pages, and they'll be reverted. You're unlikely to win a debate about this, because you can barely even cite blog posts to source your own software, and given your CV, this page is about as close to an autobiography of Sam Trenholme as the WP can be expected to have.


 * By all means take the time to improve this page. A good place to start would be to find any reliable secondary source documenting someone claiming to have deployed this software in production. --- tqbf  03:46, 5 October 2007 (UTC)

i would humbly submit, as have others, that you should cease and desist from editing this article, aside from clearing obvious vandalism. you should allow others to develop the content of this article. as the author, you have a conflict of interest. it doesn't matter how good your intentions are; you should refrain from editing this article. Anastrophe 23:39, 3 October 2007 (UTC)
 * Better him than me! --- tqbf  23:59, 3 October 2007 (UTC)


 * I agree, Anastrophe. So, please edit the article yourself to remove as many WP:SPS references as possible and reference third party sources.  Saves me from having to do the work. Samboy 14:54, 4 October 2007 (UTC)


 * IF i can find WP:RS i'll replace them. if not, i'll probably pull the material. that's how WP works. Anastrophe 16:42, 4 October 2007 (UTC)

Samboy and COI
Speaking only for myself, but the only COI issue I have with Samboy is his continuing use of Wikipedia to make unfounded claims against a competitor, Daniel J. Bernstein. Modulo that one issue, I'm comfortable leaving the article to Samboy to tend to.

For what it's worth, it's things like "staros tried djbdns and it didn't work for them, only BIND worked" that set me off. That's a total misrepresentation of the forum thread he cited, cast in a discussion about MaraDNS's own performance in such a way as to make the two projects seem on par, when performance clearly differentiates in favor of djbdns. Samboy, why don't you just stop talking about djbdns, so we can stop talking about this article?

--- tqbf 17:12, 4 October 2007 (UTC)


 * I think the problem here is that you're taking comments about dnscache personally. I mean, let's be quite frank here: Why do you get upset when I point out when people have problems with parts of djbdns.  I mean, you're (I think) not DJB, so why are you taking criticism of djbdns as if they were personal attacks?  Quite frankly, you're coming off as a DJB fanboi, who has just as much as WP:COI problem editing DJB-related articles as you claim I do.


 * As for the other point, I'd rather someone else did the editing so I could work on more important things, like the tests for MaraDNS 1.3.7. But, if noone else does the editing, I will do so. Samboy 22:03, 4 October 2007 (UTC)


 * WP:AGF. please temper your comments. the above is offensive. also, please reread WP:COI. you appear to not understand it, based upon your having cited it twice now incorrectly. Anastrophe 22:19, 4 October 2007 (UTC)




 * The words "Conflict of Interest" have meaning on WP. Read WP:COI/N: it doesn't tend to be a good sign when someone with an obvious COI accuses all their accusers of having some abstract COI. I don't know what a "fanboi" is; my relationship to these articles is that of someone with DNS security testing experience, operational experience with BIND and djbdns, and lots of questions about the notability if this open source project of yours.


 * Please also note that you didn't respond to the substance of my comment, which was that the cite you gave to the StarOS forum was misleading and involved no technical critique of Djbdns, despite the context you presented the cite in, which discussed a very serious technical issue with your own software.


 * I think you can feel comfortable editing this page to improve the content, safe in the knowledge that multiple other editors are going to ensure it stays NPOV and verifiable. You should probably strongly consider staying out of your competitors pages.  --- tqbf  03:40, 5 October 2007 (UTC)

Samboy, I'm asking you to please stop editing this article. You are obviously aware of our COI policy.

I'm sorry but if you continue to edit this article I will nominate it for deletion. AlistairMcMillan (talk) 00:04, 24 November 2007 (UTC)


 * it's been more than seven weeks since there were any edits to this article, samboy or otherwise. i think you're a little late to the table with this warning. Anastrophe (talk) 00:24, 24 November 2007 (UTC)


 * This is my fault, it wasn't intentional, and I apologize. We can probably drop it. --- tqbf  00:36, 24 November 2007 (UTC)