Re: [LUG] MAC address on Android

To: list@xxxxxxxxxxxxx

Subject: Re: [LUG] MAC address on Android

From: Tom <madtom1999@xxxxxxxxxxxxxx>

Date: Tue, 14 Jan 2014 14:34:23 +0000

Delivered-to: dclug@xxxxxxxxxxxxxxxxxxxxx

Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=googlemail.com; s=20120113; h=message-id:date:from:user-agent:mime-version:to:subject:references :in-reply-to:content-type; bh=gSKUKtbQujvT1diMYck9Dh4TLwwYm5nN6zsTDtbQMGg=; b=MUd1yxxHr1nUQCQqKxyrBFhiPoGpTierMlDAjutRKY8XUDbm/aT/j5mzlNi0VHWo+R 1JJZzDLmkhnmU8owubCICEmUjGwVS6xetodwOgNaaCxUlM9+JhtQVo91MxEGVSpVX6Nj 603vCAEDlntAlqVC7u3imHH94z7GpklIT/ZATI6yaI9CVvPrrggT5pkP2/gr5/d20IxV Wu0Z+wtr1XbFP37/+j/CRe0+RQ2wh1TvrRoZ2gzvbcxsv+PqtfyrZHsBdZZ85oZ+vXLz G8G4v0h8D/dUNrQev7SrvLUNO7ZesuiiEMK42phhTD29BnHFoxPqxm7t9nGaN6lfepGF r4sQ==

On 14/01/14 12:58, Philip Hudson wrote:

On 14 January 2014 10:09, Tom <madtom1999@xxxxxxxxxxxxxx> wrote:

TCP/IP is a rubbish tech

You said it. As soon as I posted my last I thought "I should have said DNS and DHCP and NAT are rubbish tech" then I thought "Might as well just say TCP/IP" (though I think TCP itself is kinda cool).

- but it works.

As I always say when this phrase is presented unsupported as a definitive answer:

... for some value of "works".

In other words, you're only restating the problem. "Works" in this context has the same pleasing property that weasel words like "natural" and "normal" have in other contexts: namely, they mean exactly what you privately intend them to mean, rather than what anyone else has agreed to understand by them, which is always good fun. No use at all for communicating meaning and resolving problems, though.

Or, as someone else put it, the only possible agreed meaning of "works" in this formulation is such a degraded one -- like it compiles and runs without aborting and as documented -- as to be meaningless and useless. The alternative to this "it works" is not some software (or a protocol stack) that "doesn't work", it's a pile of random bits.

The real problem with the TCP/IP stack as I see it (and I don't claim any originality of insight here) is that it was primarily designed for primarily static networks administered by people who understand and like octet triplet quartets (with masks!), but is used almost exclusively for dynamic networks by people who don't know their bytes from their bits. The result is a living hell of leaky abstractions and real-world impedance mismatch, opacity, incomprehension and blockage. The AppleTalk stack was designed for dynamic networks and non-techie end users. It's a great pity that we ended up where we have. GNU could and should have tried to do better.

I'm not clear on just how well, if at all, IPv6 will resolve these issues. Anyone care to opine? Is there hope?

--
Phil Hudson http://hudson-it.no-ip.biz
@UWascalWabbit PGP/GnuPG ID: 0x887DCA63

Time to configure devices on networks is considerably less than getting the hardware out of the box and plugged in so there is no need for hope. Where we need hope is that the users who are half educated or just pointlessly demanding can be shut up with an upper cut. Funnily enough I've found its the people who think they need a Mac for browsing the internet cause several orders of magnitude more handling than learning octal in several dialects of Aramaic while drunk.
Tom te tom te tom