[Outages-discussion] [outages] Comcast Issues Central & East

[Jay Ashworth]

Or, the short version of my favorite design precept: "get the glue right".

This can be as trivial as if you have video to push out of a device, use analog VGA unless there is a compelling reason not to.

On August 1, 2019 2:29:24 PM EDT, Jeremy Chadwick <jdc at koitsu.org> wrote:
>On Thu, Aug 01, 2019 at 01:02:37PM -0500, Eric Spaeth wrote:
>> On Wed, Jul 31, 2019 at 2:11 PM Scott Weeks <surfer at mauigateway.com>
>wrote:
>> 
>> >
>> > ----------------------------------------------
>> > and uses RIPv2 to announce its LAN-side IPv4 block
>> > ----------------------------------------------
>> >
>> > They use what!?!?!  :)
>> >
>> >
>> The right protocol for the job.
>
>I suspect the "they use what?!?!" comment is driven by the age-old
>mentality of "RIPv1/v2?  OMG who uses that, nobody uses that, it's old,
>limited hop count, it uses multicast, it's terrible, it should be
>replaced then taken out back and shot!"
>
>This overall belief is a very broken one, yet continues to rise/bloat
>in
>the day and age where everyone wants "new shiny" and thinks "newer is
>better".  I've seen similar reactions to all sorts of late
>ARPA/DARPA-era protocols, even ones as "basic" (except not really) as
>telnet.  Rarely is good/solid technical justification given for this,
>instead it seems to be driven by dislike for having to read RFCs and
>disapproval towards all things "old".
>
>More rarely is consideration given to the fact that something is used
>because it is technically better/more relevant to the situation (such
>as
>here), or because it's substantially more well-understood (this latter
>point is often overlooked.  Example: syslog has some shortcomings (see:
>RELP wrapper), but it is a well-understood protocol and is a lot easier
>to troubleshoot comparatively to things like, say, logstash).  I am not
>implying alternate or newer routing protocols are not well-understood,
>I
>am speaking about the mindset.
>
>Picking the right tool (protocol) for the job is the proper approach,
>no
>matter how "old" it may be.  I wish more people asked "why was {thing}
>chosen and when?" first.  The why, as demonstrated here, is important.
>Because newer is not always better.
>
></soapbox>
>
>-- 
>| Jeremy Chadwick                                 jdc at koitsu.org |
>| UNIX Systems Administrator                      PGP 0x2A389531 |
>| Making life hard for others since 1977.                        |
>
>_______________________________________________
>Outages-discussion mailing list
>Outages-discussion at outages.org
>https://puck.nether.net/mailman/listinfo/outages-discussion

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/outages-discussion/attachments/20190801/0e7a677a/attachment.htm>