lsmod | grep stupiduser

Courtesy Lyza Danger Gardner

It all started so simply.

It occurred to my yesterday that it might be nice to be able to conveniently lay my hands on some local data and media while away from home.  Sounds simple enough, right?  That’s what SSH is for.  So this morning, over coffee, I boot up the big beige box that’s been lurking, abandoned, under the spare desk in the office.  Immediately, my sensibilities are affronted:  that crufty old Fedora Core install was so two years ago.  It had to go.  Besides, I wanted lean and mean for this application, something that could run five nines on hardware that wheezes audibly even when powered off.

Twenty minutes and another cup of coffee later, Ubuntu Server 8.10 is settling into its cozy new home.  Now what?  Well, SSH first, naturally.  After all, what fun is server administration if you have to sit in front of the actual server the whole time?  Then came Fuse/sshfs, AutoFS, and SAMBA.  Then my old buddy Vim and I put in some quality time together, and, when the dervish of config files had subsided, we were cooking with gas.

Sort of.

Except that two of the Windows machines wouldn’t talk to the server, and one of them was invisible to every other host on the network, despite the fact that they’d all been chatting happily just the day before.  After what was certainly no more than 45 minutes and fifteen reboots, everyone was playing nicely, and it was freshly mapped network drives all around.  Then I was free to focus on the actual ‘remote accessibility’ aspect of this remotely accessible server.

Have I mentioned that it was nearly 11:00 AM now?

At this point, ominous phrases like ‘dynamic IP’, ‘NAT’, and ‘multiple virtual hosts behind two different routers’ began to chime in my head.  Fortunately, there was one cuppa left, and I was able to steady my nerves long enough to not only acquire a DynDNS account, but to also invoke enough iptables voodoo to make everything map reasonably sanely.  If you squint a little.  Not, of course, that I could tell without first hopping on the neighbor’s WiFi, because I failed my Diplomacy check and couldn’t convince the external router to let me reach the WAN address from inside the network.

So now I’m on a roll.  Over my turkey sandwich, I start thinking:  Ok, so, SSH, scp, etc, etc are good, and they’ll do for quick-and-dirty operations.  But–and call me spoiled if you must–I like a little polish in my workflow.  I’ve got a spankin’ new network back in the fortress of solitude, and meanwhile me and my trusty portable dual-core steed are riding the lonesome trail.  What I’d like to do is tunnel back into the local network and, via SAMBA, keep my drives mounted just like they were when I left home.  Well, it’s fine and good to do that–to simply tunnel ports 139 and 445, if I’m on a Windows box–and let it go at that.  The problem is that that b0rks Windows filesharing over any other peer-to-peer net you might be on.  Like, say, the one in the lab.  That I use all the time.

So I do some googling, and the first three or four promising hits mention things like installing virtual network interfaces.  Yawn.  Then it hits me that I’ve got a JeOS server appliance running in the background on the laptop anyway.  Could I maybe…

Nah.  That’s too crazy, even for me.

But out of sheer morbid curiosity, I ask the internets.

Sure enough, I’m not the only one to have thought of it.

Remember the sshfs install earlier?  That’s going to come into play here.  What I decided to do is this:  I use Fuse on the JeOS appliance to remotely (and securely) tunnel to and mount the germane drives on the gateway server, which are in turn simple CIFS mounts of the same network resources I would have access to behind the firewall.  Then, I install another SAMBA server–again on the JeOS device–that will share the net-mounted drives with the host laptop,

When you stop laughing, I’ll go on.

Ideally, I would then fall back on some Powershell scripting goodness on the Windows side and some preexec/postexec handwavery Linux-ward to make the the mounting and transitioning as seamless and transparent as possible.

Ideally.

In reality, what happened was that everything worked fine up until I had to reboot JeOS a few times and there somehow came to be issues with the known_hosts file, making SSH panic comically.  Then there were problems with NTLM and login-mode compatibility.  Then I found out the fun way that running a VMWare appliance in NAT mode–ie, with both of those clever virtual network interfaces running–means that Vista is bloody well determined that you are on an unidentifiable netowork, say what you will.  This means, of course, that the ever-helpful firewall slams shut, and alas, no filesharing.  Two quick registry tweaks later, and that was more or less straightened out.

…by which time it was nearly 5:00, and I had completely forgotten just why it was that I wanted remote access to my network anyway.

Home networking:  my anti-job.

**EDIT**

I realized as I was falling asleep last night that VMWare Workstation has guest-to-host virtual drive sharing capabilities.  So much for that extra SAMBA server.

We Are Web

In case anyone missed it, Andy Borowitz published a cute, Onion-esque piece in the Huffington Post yesterday jabbing at John McCain and his much-publicised complete lack of familiarity with post-Reagan-era technology.  I don’t really want to talk about that piece.  I mean, it’s funny.  But seriously, can we all, just for simplicity’s sake, agree on a couple of points here?  Is McCain old?  Yes.  Is he completely, ludicrously out of touch with modern society?  At least insofar as that society is concerned with computers and the Internet, by his own admission, he is.  Is it inadvisable–perhaps even dangerously irresponsible–to even consider appointing a man who lacks even the most rudimentary understanding of the technological (and indeed social and cultural) phenomenon which now encompasses and in some ways defines a great part of human civilization to the Presidency of the United States of America?  Emphatically yes. (Don’t you love run-on sentences?)  Now that that’s settled, I say again, I don’t really want to talk about Borowitz’s article.

No, what I find more amusing, and infinitely more thought-provoking, is the minor comment storm over on FriendFeed that arose around a link to said article posted by Kevin Fox (and continued in a later cluster spawned by Robert Scoble).  It’s kind of a windy one, but I picked out a couple of major points:

From Flickr, courtesy Mathias Pastwa

Inclusivity and Social Networking Etiquette

This whole interchange started because one user–Fox–tried to initiate a private (or at least, semi-private) conversation amongst a close circle of friends, while another user–Scoble–wanted to spread the discussion around as much as possible.  The topic of social networking etiquette is getting a lot of mileage right now, so I’ll leave speculation on the who-follows-whom, clique politics as an exercise for the reader.  The takeaway here is about something more fundamental, anyway:  Can I, or should I, shout every conversation into the cloud just because it’s there?  In other words, is it ok to constrain distributed discussions?

The answer, plainly, must be yes.  I wouldn’t stand in the middle of a crowded bar and bellow at large words that I meant to be heard by someone sitting next to me.  This would be true even if every person in the room was my best friend.  Inevitably, if the user experience is going to remain a positive one, FriendFeed et al are going to have to institute some means of scope control.  Consider, for instance, how I came to be privy to the conversation in the first place:

I follow Robert Scoble (who I know is positively giddy to know that he can number me among his n-thousand netstalkers);  Scoble interacts with Kevin Fox; FriendFeed’s arcane Friend-of-a-Friend voodoo kicks in, and, bam, Fox’s post hits the top of my page.  Now, Kevin Fox was not talking to me.  Kevin Fox does not know me, and does not care what I think.  He did not want my opinion (and said as much).  But there I was, friend-of-a-friend of no one in particular, a mere hanger-on, listening in by virtue of the fact that I can click a button by an A-lister’s name.  It is a problem.  Because I can’t listen quietly.  Oh no.  I have opinions.  Lots of them.  And so does everyone else who clicked that same button I did.

With me and 30,000 of my closest friends breathing down the same dozen or so necks, we’re looking at some fairly low-SNR conversations.

So, again, scope control.  Unfortunately, I can’t just check a magic box to restrict a post to My Friends until I can define who they are.  Now, that could be as simple as defining friends as those followers who you are following reciprocally.  Or maybe we should implement a Facebook-style system–which I think of like those security doors which require two keys to open:  a user has to friend, and be friended back, to get a spot at the table.  Then again, we’ve already got Facebook, which is starting to look more and more like FriendFeed, anyway.  (Or vice versa.  Whatever.)  And let’s face it:  that sort of system virtually guarantees that grubby, unwashed, pit-standing peons like me no longer even get to window shop outside the halls of power.  Obviously not my first choice.  So perhaps we should seek to model our online relationships after our real-life ones, segregating our sundry online contacts into graduated spheres.  The innermost, a user’s actual close, personal friends, would be at the center, and would almost certainly fall under a mutual-aknowledgement style of authentication.  Outside this, a slightly larger group:  the people you work with, blog about, perhaps even know on a first-name basis, but don’t necessarily get Labor Day barbecue invites from.  And finally, the outermost circle:  a loose confederation of fans, admirers, stalkers, wannabes, sycophants, and people looking to glom some influence or popularity from you.  Then constraining online discussions–even on massively-distributed, cloud-spanning systems like FriendFeed–really would be as simple as ticking a checkbox.

I don’t like this for a couple of reasons.  First, it sounds like a lot of administrative work for me.  (Well, not me, specifically.  Nobody on the Web knows or cares who I am; I’m talking about me as in the hypothetical, well-connected personage.)  Frankly, having friends should be as low-effort as possible.  Secondly–and here I’ll ask you to forgive the impertinence of an upstart with a capital ‘Newb’–that’s not what the system is designed for.

Distributed comment/discussion networks like FriendFeed are, and will continue to be, attractive primarily because, all else aside, they do one thing very, very well:  they dissolve the rigid, linear system represented by concentric spheres of connection, influence, and ideas into a coherent, self-propagating thoughtspace.

And isn’t that the just what we’ve pined for since the early days of the Web?

I don’t mean to make light of crowded-room issue; I know that not every conversation is intended for digestion by the collective.  But there are options.  FF has Rooms, as many users pointed out in the original comment threads.  And there are always direct tweets, instant messages, emails, etc, etc.

Sometimes, if the bar gets too crowded, you have to take your conversation outside.

From TechCrunch, thanks to Scott Beale/Laughing Squid

Ownership of Discourse

A couple of comments arose in Fox’s cluster that centered around the notion of propriety of a discussion.  In particular, there was expressed some concern over the implications of ’siphoning off’ commenters from another person’s thread/cluster.

*sigh*

Unfortunately, this is a tough nut to crack.  The bottom line (literally) is that page views are the bread and butter of bloggers, and comment streams drive page views.  The old wisdom, then, would say that anything that takes eyes and minds offpage–that is, away from a blog post–is tantamount to picking the author’s pocket.  And since, by definition, all comment clusters across services like Disqus, Intense Debate, and FriendFeed occur offpage, all such services must be bad for (the monetization of) blogging, right?

Well, maybe.  The fact is, there’s just nowhere near enough data to say for sure yet.  Personally, I believe it’s a moot point.  Why?  For the simple fact that, 15 years ago, very few people would have been able to conceive of men and women being able to earn a sustainable living as bloggers.  There was no imaginable economy behind internet journalism and commentary in that age.  But the Web is about adaptation.  Monetization models change.  Bright men and women with far greater knowledge of economics that I will find a way to parley my raging information addiction into cash in the pockets of the likes of Scoble and Louis Gray and Dave Winer.

After all, what is a blog but a static interface to an author’s content stream?  FriendFeed and other ‘lifestream’ systems simply turn that interface into a dynamic one.  This means, basically, that a content stream–opinions, ideas, information–is attached to a person, an author, and not a URL.  When all discussions are held cross-cloud, there’s no such thing as hijacking a discussion because the discussion is everwhere. Then the Web becomes–finally–about ideas instead of hitcounts.

In a nutshell, what I’m saying is this:

What people think–and how many people think–about what you have to say is infinitely more important than how many people visit your website.

Every idea, every thought, is content now.  Pages and threads are yesterday’s notions.  Advertisers and monetizers will catch on.  They always do.  It means change, yes; and change is always easier to swallow for those–like me–with little invested in the current system.

But change will happen one way or another.