Yup; the Internet is screwed up.

Michael K. Smith - Adhost mksmith at adhost.com
Wed Jun 22 20:19:13 UTC 2011

On 6/22/11 12:48 PM, "Jeroen van Aart" <jeroen at mompl.net> wrote:

>Steven Bellovin wrote:
>> When I was in grad school, the director of the computer center (remember
>> those) felt that there was no need for 1200 bps modems -- 300 bps was
>> fine, since no one could read the scrolling output any faster than that
>> anyway.
>> Right now, I'm running an rsync job to back up my laptop's hard drive
>>to my
>> office.  I hope it finishes before I leave today for Denver.
>I understand the sentiment, but the comparison is flawed in my opinion.
>The speeds back then were barely any faster than you could type, I know
>all too well the horrors of 1200/75 baud connectivity.
>Luckily nowadays now it's about getting your dvd torrent downloaded in 2
>minutes, vs. 20 minutes, or 2 hours. Or your whole disk backed up before
>your flight leaves. You're now able to back it up online to begin with.
>The thing here is that I talk about *necessity*. Once connectivity has
>reached a certain speed threshold having increased speed generally
>starts leaning towards *would be nice* instead of *must*.
>And so far the examples people gave are almost all more in the realm of
>luxury problems than problems that hinder your life in fundamental ways.
>If you have a 100 mbps broadband connection and your toddlers are
>slowing down your video conference call with your boss by watching the
>newest Dexter (hah!). Then your *need* can be easily satisfied by
>telling your toddlers to cut the crap for a while. Sure it'd be nice if
>your toddlers could watch Dexter kill another victim whilst you were
>having a smooth video conference talk with your boss, but it's not

To paraphrase Randy Bush - I hope all my competitors work on their version
of what their customers "need" versus what they "want".  Why on earth
would you not want to give them what they want?  Why does "need" have
anything to do with it, particularly when "need" is impossible to quantify?


More information about the NANOG mailing list