Archive for April, 2008

Pet Peeves with New Laptop Screens

Posted in Laws by epictetus on April 28th, 2008

Why, oh why, are laptop manufacturers making their products worse and worse? I’m talking about screens here — two of the major changes in the past few years to laptop screens are both simply awful:

  • Glossy screens. If you wanted a mirror, you would buy a mirror, not a laptop. These make the screen VERY hard to read because you’re staring at a reflection of your own face (and the room behind you). I have read some other people who agree with this; their general idea was that glossy screens look shiny and new when someone goes to Best Buy and looks over all the laptops on display. I actually returned one notebook because had a glossy screen; as far as I’m concerned, a glossy screen makes a notebook essentially unusable.
  • Widescreen aspect ratio. I don’t know about you, but the vast majority of the documents I read (or work on) are not very wide — they are TALL. That’s why mice have up-down scrollwheels and not left-right scrollwheels — because most of the time documents are _vertically_ too large for the screen they are on, not _horizontally_. The only legitimate reason to use a widescreen aspect ratio is for watching widescreen movies — I don’t watch movies on my laptop. If anything, I would like to buy a tallscreen notebook so that I can spend less time scrolling. Widescreen aspect ratios also cause a lot of problems including:
    • Looking horribly stretched and blurred whenever something (a movie, a game, etc) needs a normal 4:3 aspect ratio.
    • Incompatibilities with virtual machine software
    • Incompatibilities with different operating systems
    • Incompatibilities or just plain ugliness when using Remote Desktop or similar between widescreen and non-widescreen PCs

Why Web-Based “Easy DNS Checker” Tools Cannot Reliably Troubleshoot Reverse DNS bounces

Posted in Laws by epictetus on April 25th, 2008

DNS can be somewhat complicated — SOAs, glue records, MX records, PTR records, registrars, CNAMEs, etc.  It’s a lot to learn, and a lot of engineers don’t understand the tools (whois, nslookup/host, and dig being the three most important ones).  I’ve seen a lot of issues recently when people are trying to configure mailservers, and having trouble sending to certain domains (getting bounce messages related to reverse DNS).

So, the engineer visits or one of these other awful sites, plugs in the domain name, and gets printed out a 12-page report giving them a warning about their mailserver’s reverse DNS and some suggestions on how to fix it.

However, in many many cases, this report is wrong, and here’s why:  The way reverse DNS checks typically work, when the mailserver makes an outgoing connection to a foreign server (to deliver a message), the foreign server checks for a PTR/RDNS hostname record associated with the IP your mailserver is connecting from.  It then does a forward dns lookup on this hostname and tries to confirm that it resolves to the same IP. If it does, you pass the test. There is absolutely no way, whatsoever, that a web-based DNS tool could possibly know what IP your mailserver makes outgoing connections on.  What most of these tools do is check your MX records and assume that the IPs that you receive incoming mail on (from the MX records) are the same as the IPs you send outgoing mail from.  Many people these days use incoming mail filters that are not necessarily on the same IP addresses as their outgoing mail server.
The right way to test this is to do the same test manually — first, figure out what external IP your mailserver makes outgoing connections on (by hitting, for example).  Then, use nslookup or host to figure out if there is a reverse DNS record for this IP.  If there isn’t, that is your problem right there.  If there is, then use nslookup again to check the _forward_ lookup on that same hostname — this must match your outgoing connection IP.

View of Boston from Boston Rock, Prospect Hill, Waltham

Posted in Laws by epictetus on April 2nd, 2008

Boston Rock view