[clue-talk] A Significantly Better Mousetrap, was Re: [clue-tech] Mephis linux

[Nate Duehr]

David L. Willson wrote:
> Unless I misread your post, you're unexcited by any of the things others are doing, and you have no exciting ideas of your own to promote.  I'm sorry, but I find myself unable and unwilling to contribute to that mindset, and I find it darned curious that you would think the perspective valuable enough to share considering how unlikely it is to motivate anyone to try to do anything.

Heh... I have plenty of "exciting ideas" but they are unrelated to Linux.

Usually they're in the form of integrating RF systems for Ham Radio and 
others.  Linux is sometimes used, but it's not critical to any of it.

I'm a USER of computers, not a developer.  Sometimes you just have to 
know your own limitations.  Just like I'm not a politician, I still can 
have opinions about how they do their thing -- I have opinions about 
software development.

I also work for a company that develops a lot of software, and have been 
doing support for many generations of the stuff... so my perspective is 
different than many Linux fans who are almost always developers of some 
sort.

(I've actually been fairly careful to keep specifics about my 
experiences with corporate software developers out of this thread, but 
they certainly color my opinion.  Without getting into too much detail, 
there have been a lot of promises that "oh, we'll fix that in the next 
release... or in the next hardware platform" of core problems on things 
I've worked on at every company I've worked at for a couple of decades. 
   Those fixes either are botched (happens more often than companies 
want to admit) or they're de-scoped by managers... and even though the 
support staff knows the problem has been around and annoying customers 
for decades, it's still there... in every version released... but, I 
also see the same types of things going on in open-source.  No difference.)

The most development I do is shell scripts, most of the time.  I get a 
wild hair once in a while and write something in another language other 
than the shell or Perl, but not often.  (No need to, really.)

By many developers standards that means I have no say in what they work 
on.  That's fine, but Linux zealouts always say that "the problems get 
fixed by many eyes"... no they don't.  They get fixed if the devs want 
them fixed.  The users are rarely consulted on what they'd like to see 
fixed.  You can turn in bug reports and beg as much as you want, unless 
a dev decides your problem is worthy of their time, hundreds of people 
can complain and the bugs will remain open from release to release.

Having worked in customer support for a long time, I get it that some 
bugs are minor and stupid, but I also believe that if a customer says 
it's a bug... it's big enough to them to want it fixed.  No matter how 
"stupid" the bug is to a dev or their managers.

In the case of Linux, there's not even a manager to "plead" the case to. 
  You plead to the dev/package maintainer, and if you don't convince 
them... it'll never get fixed.  At least when you buy the software you 
have a LITTLE clout... maybe not much these days, since software is 
virtually a commodity now, other than big important company-running 
types of software, and those packages are so big, only the "IT 
Department" can tell the vendor what they want fixed and in what (slow) 
timeline... end user input again, is invalidated by the process of going 
through multiple layers to get to the dev.

So in all... I'm just curious how we get to a world where the end-users' 
desires trump all and actually get worked on.  Aren't the end-users the 
experts on how they want the computer to behave?  The devs often know 
how to get the computer to do that, but find the solution as "too 
mundane" or whatever... and often don't even try to fix the problem.

That's been my experience anyway...

Linux isn't "broken" in the sense that it doesn't work on the desktop. 
Linux just has lots of little cracks and busted little stuff, making it 
kinda rickety on the desktop.  It's like a house built just a little to 
fast by the framing crew, even though they had no real deadline.

(What's the hurry to release new code for Linux anyway, other than to 
keep up with or get ahead of where commercial software is at?)

> The one thing you seem a little interested in is the "real fix" spam problem.  Go design that fix.  Then, when you're all done designing and beginning to implement, tell me again how Free vs. Proprietary is not a meaningful debate.

The design is done (by others) and has been available in the form of 
RFCs for years, it's the implementation that is stalled, since it 
requires a trusted centralized clearing house and large organizations to 
start wanting only authenticated servers to talk to their servers, 
pretty much overnight.  Requiring all servers authenticate to one 
another with real SSL keys (TLS), handles spam very well because it 
keeps the problem in the SOCIAL realm... you can now FIND the offending 
server.

Once you can ALWAYS track back the source to a specific server you now 
have the power to block the morons sending spam and only those morons 
without guessing.  An RBL based on the SSL keys of the servers sending 
bad mail becomes ultra-powerful.

It also adds an economic incentive to run a clean server, since your 
"reputation" goes along with your SSL key to every other server you 
visit, and getting a real key is (relatively) expensive for individuals 
and small spammers, and big spammers alike.  Etc.

The spam problem is one of politics, not technology.  Tech won't fix it 
until the political motivation is high enough.  It probably won't ever be.

Spam's detrimental, but not detrimental enough to have a large company 
(or better a government) stop accepting standard un-authenticated SMTP 
connections one day (even with plenty of warning), and possibly block 
communication from a customer.  They'd rather spend thousands of dollars 
on spam "guessing" engines instead.

I have to admit, you just made me think of some really good technology 
that WAS implemented first on open-source... SpamAssassin.  Very useful 
stuff, and an example of something that does NOT have corporate coders 
working on it!  Cool.  That was innovation.  It does happen, after all.

Nate