Open Letter to the Webmaster at NASA

July 15, 2009 By: erik Category: Complaining, Geeky, Internet 342 views

Rate this post:
1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)

Dear NASA Webmaster,

Your organization has been using the internet for at least 20 years, way back before there existed a World Wide Web or even a Hypertext Transfer Protocol. Even I remember the days when Gopher was the most user friendly internet service. In the intervening years, however, you may have noticed that a few unwritten standards have cropped up.

Principally, I’m referring to the total dominance of HTTP and HTML as the primary method of communicating between servers and human clients on the internet, and how every single site on the internet will forward you to if you attempt to visit without the now-completely-assumable www prefix. Some websites, in the interest of brevity, will forward you the other way, preferring their website to not use www. But the important thing is that a web page will show up with or without the www!!!

Every time I visit your website, I type in into my browser’s address bar, and every time, it fails!! To actually find your website, I am forced to hit my W key three times and go to I know of no other websites in 2009 where there is a distinction made between and Please, please, please fix this!

Erik Rasmussen

This letter has been submitted via NASA’s contact form.

Update! Within an hour of me submitting this, and before this post was published, I got a response from NASA:

If you’re using Safari or Firefox you can type in and it will direct you to NASA’s Web page. For other Web browser’s please use

This is, of course, not true, as I am using Firefox and visiting fails. While normally, I have no problem with ignoring non-Safari-or-Firefox users, fixing this problem would take, at maximum, three minutes for anyone who knew what they were doing with DNS or Apache. Boo, NASA!

  • Ray Tibbitts
  • jane

    Yeah, good luck.

  • Paul

    For me, both Safari and Firefox send “” to “” quite satisfactorily.

  • Josh

    Hmmm, doesn’t work for me in IE8, Firefox, Safari or Opera. Lynx, on the other hand, doesn’t have any problem resolving. If you’re looking to make a collection of similar sites, check out

    News Flash! I think we’ve jumped the gun on this one. Ray, Erik and I all failed to resolve to At least Ray and I are using Opendns (and Erik is geeky enough that it wouldn’t surprise me if he uses a similar service.)

    However, Paul has no problem on his system, and when I fire up a secondary box to try with lynx, I don’t have a resolution problem either. Hmmm, strange. As a double check, I open Firefox on this box and it too resolves without a problem.

    Just for a laugh, I take a look at my list of DNS servers on each box, and discover that while I use Opendns exclusively on my main machine, the older one has a remnant of a previous setup — a reference to one of Telefonica’s old DNS servers, in addition to those of Opnedns. When I comment out that address, I can no longer resolve to

    My initial guess was that the error was Opendns’, but doing some “diging” I’m starting to think that things are the other way around. NASA has no CNAME record, and it looks like the old Telefonica server hasn’t been patched for the whole Kaminsky thing. No definitive answer yet, but I’m getting closer…

    Will become a litmus test for vulnerable servers? More later, gotta dash.

  • Erik R.

    Both the machines I tested this on, one in Spain and one in the US, were using OpenDNS. I bet the problem is that your browser is smart enough to try prepending the www if and only if you get a “I dunno!” response from DNS. But since OpenDNS tries to be more helpful than that, the browser’s fallback is never triggered.

    While I still maintain that this is a failure of compliance on NASA’s side, perhaps OpenDNS could add a “check with www” test just in case.

    Thanks for all that sleuthing, Josh!

  • simon doesn’t work for me.
    Although, to be fair, I still fairly regularly come across other sites with the same problem. Pí¯sses me off too.

  • Josh

    Actually, thinking things through more clearly, I think that I jumped to conclusions. My guess is that we’re seeing something similar to the “Case of the 500 mile email” ( When I query NASA’s dns server for a CNAME record, Telefonica eventually hits (3rd NS on the list) and resolves to the IP. However, the response time is somewhere between 800 and 1000ms. Opendns only takes between 200-400ms to resolve, but doesn’t find the CNAME record. Looks like one of several things is happening:

    1. NASA uses an old, pokey server for their tertiary NS. (In this case, they should consider either upgrading this segment of the network, and/or putting aliases on one of the newer, and presumably faster, machines.)

    2. Opendns has a shorter timeout than Telefonica. The advantage to a short timeout is a faster go/no go resolution to a query. (In other words, queries either open in a browser or tell you quickly about the failure.) The disadvantage is in cases of poorly configured/maintained NS like the current situation.

    3. Hypothetically, Opendns is configured to limit queries to the first 2 NS before passing the buck on to a higher reference. (Is this even possible? Any BIND gurus out there?) The problem here is that the buck stops at NASA’s servers.

    Any other options?

    I’d be interested in other peoples response times. I’ve used “$dig @ -t CNAME” and “$dig @ -t CNAME” for Opendns and Telefonica.