Traffic boost

Well, sort of.  My RSS feed attracted a fairly sizeable number of hits in the last 24 hours, compared to the rest of my website ;).  Hardly unique, of course, but still.  It’s been live for all of a day, and has attracted about 1/5th of the traffic IN A DAY that the rest of my site has accrued in a month.  That’s hits speaking, not bandwidth, but still!  Mildly insane.

I can now see why larger sites with RSS feeds request you keep your updates to an hourly timing cycle or whatever!  Not that I want anyone to do that… I don’t really care, and I’d imagine that, at 3KB a pageload, Dale’s connection doesn’t really, either.

So, when I was looking at my logs, trying to figure out what kind of traffic impact the RSS had had, I was looking at the clients people were using to access RSS.  Awstats doesn’t identify more obscure User-Agent strings (despite reputedly having 97 browsers which it supports), so I looked under “Unknown”.  Which was cool.  A viewer called “Soup” came out on top, then KNewsTicker (I run that one in my panel, and updates are set to a meagre 5 minutes!  Should probably set that value a little higher…), then FeedValidator.

Which is all well and good.  But then I see another one in there.  “Mediapartners-Google/2.1” – a few searches later, it’s revealed that this bot is used to spider websites using AdSense…  gee, thanks guys.  Nice to know that you think my measly 200 unique impressions a month is worth advertising to, really…

Okay, so now I’m really curious.  Why am I being spidered for a service I’m not even advertising with?!  Seems kind of stupid to me.  Seeing this hosting isn’t costing me anything, I’ve got no reason to try and recoup costs on it… and really, that’d be all I could ever hope to achieve with Google Advertising, unless I saw a few hundred-fold increase in traffic.  Which, of course, would probably mean I’d need to start paying for hosting somewhere with a nice amount of bandwidth and resources (got to love working for a webhost, at-cost hosting is a nice perk)… that said, it’d make more sense to purchase a TLD first.  No, a frame-based redirect from some pacific island doesn’t count, as cool as Tokelau must be.  At any rate, that kind of traffic seems to be a few months away yet. (Yeah, web expansion is the one area I’m an optimist… it WILL be months, not years ;) )

Has anyone else who is a bit pedantic about their log-checking noticed anything like this?  I don’t care if you’re using AdWords on your site – that’s justifiable – but if you’re NOT, and still getting this weird spidering… yeah, if you’ve got any ideas, let me know.

Now with RSS

Not that I’d ever be one to follow a trend, but you know… these things happen.  I’ve got an RSS feed running @ http://sc.dalegroup.net/rss-2.0.php for all who are interested.

I’ve taken an “embrace and extend” approach to it, though – I’m hoping that those I was influenced by in implementing RSS will take note of the new feature I’ve added, and enhance theirs accordingly!  In addition to feeding the title of the post, and the first 128 characters (rounded up to the nearest complete word) of the body, my feed also denotes the number of comments.

Not that that is a particularly useful feature on a website such as this, though – I get hardly any comments, which is sad :(  But hey!  That’s not what this is about… I’m hoping other people will also add this to their RSS title so that I can see at a glance in Ximian if people have commented on a particular post.

Enjoy… comment more… speak Dutch…

World’s biggest system monitor?

My current project (as of a few days ago, nothing long term ;)) is getting a working phpSysInfo page with trippy temperature monitoring and other such kah-razy features.

So, I grabbed the latest from the SF CVS server (2.3-cvs) and installed it, because my old version (2.2-release) has some serious issues with SuSE’s way of doing things – it “worked”, sans Memory Usage and missing much of the Hardware Information.  Not that that is really relevant anyway, seeing I (of course!) wanted to try out the latest.

I’m not sure if 2.2 supported the trippy hardware monitoring thing, but even if it did, I’m over it already ;)  2.3 has a nicer version number :p

Yeah.  So.  Hardware monitoring.  I downloaded a few (it supports 4 different backend programs) and tried to compile – xmbmon downloaded and compiled fine (well, okay – mbmon compiled fine, the x extension didn’t… not that it matters, because for my purposes I only want the CLI version)… except it’ll fail except when run as root.  It’s a documented problem, although the only reference to it was in relation to *BSD systems, and the fix refers to some kernel-related file which apparently doesn’t exist.

So I gave up on that… it was probably a surmountable problem, but still, other peoples code scares me off.  Moving on to the next (non-BSD-only) option!

LM Sensors… hey, that’s okay.  Relies on kernel hooks, which prior to 2.6 kernels involved rolling your own with an i2c extension compiled in.  I am, for the first time, as greatful as I should be for SuSE’s lean towards the cutting-edge!  Hmm.  That said, LM Sensors apparently won’t compile without kernel source.

So.  I want to install a 780KB app, and wind up downloading ~700MB of stuff!  Hmm.  This works, really it does.  Kernel sources are only ~180MB (at least, the SuSE respiritory RPM’s are that big… last time I checked the size of the kernel (admittedly, that was back when 2.4 was the new thing), it was about 60MB!), but I got distracted in package-selection, and saw that a newer version of Opera was available, so I grabbed that.

Apparently the old version of Opera had no problems at all without a certain dependency, but this latest one requires Eclipse… a ~170MB Java library thingo.  At least, I think that’s what it was… OSS is way too trusting with dependencies!  Hehe.

So after having downloaded all that, I’m thinking the compile still isn’t going to work!  Doh!

Ah well.  phpSysInfo is still cool ;)

Woo! Now with system monitoring!

Yay!  Check it out @ http://nicktangents.is-a-geek.com/info/phpsysinfo-dev/ whenever my computer is turned on.  It was all easy once I read the QUICKSTART file, hehe.  Yes, I did still need the kernel sources.

Was just a matter of compiling, installing, running sensors-detect, senors -s and copying a few files so it does stuff at boot time!  Cool program, very cool.  If you’re a Linux-ite, check out http://www2.lm-sensors.nu/~lm78/ for yourself.  Worth having ;)

J.

RTFHTMLD!

I’ve been hacking a JavaScript dropdown menu for a client’s site on and off over the last few days, but only finally sat down with a view to fixing the crazy thing this evening. My goodness. You know those moments where you kick yourself so hard you pity the chair you were sitting in? Okay, so I sit in front of my computer too much, but you get the point. This is like… ouch. ARRRRRRRGGGGGGGGHHHHHHHH!!!! Calm is good.

A note, for anyone who is looking to implement not-overly-documented dropdown menus – read your base source file in its entirety before even contemplating trying to implement. It was two lines. Two bloody lines of JavaScript which turned into a ~1 hour nightmare search – not to mention time which I haven’t counted, because it was hackish, not goal-oriented coding. Of course, the longer spent, the more it hurts to find out exactly what I’d missed.

  var ddmx = new DropDownMenuX('topnav');
  ddmx.init();

Those two lines. So much pain!

Oh, and just to clear up any confusion, the acronym which is the title of this news post ends in D, for “Document”. Not “dotcom” or anything stupid like that ;)