A quick week

Spent the week at SANS Virginia Beach.  It went quickly.  SANS chose a better hotel this time around (it's been awhile since I last went to one), with the break food/coffee being better and having free valet parking for local attendees.  I attended the "Virtualization and Private Cloud Security" course, taught by Paul Henry.

Ran into some old friends (hadn't seen Steve S. in a decade) and made some new ones.  Got to show off some of my tools and generated some ideas for new features (spent the last couple hours, before posting this, adding bulk processing to the document manager).  Overall, the week was engaging/interesting.

Concerning the document manager: expect a rewrite in the near future.  I've tacked on/edited enough features that the code is beginning to look a bit messy and at least one function doesn't align well with the overall schema of the program.  It works will enough to suit me but will annoy anyone else using the software.  It's on my "to do" list but not exactly near the top.

The next few weeks will be a bit busy:

• I'll be re-certifying my CEH next month (thanks go to my supervisor!).  I'm currently certified under v5 and need to brush up on a number of things.  Time to hit the books.
• The local hamfest is coming up fast (early September).  I have a couple things on the "must find" list that can't wait until the January FrostFest.  I managed to miss Field Day again this year so some of my projects are still on hold.  Want to attempt Extra Class this year.  Again, time to hit the books.
• There's probably more eye surgery pending (it's still a bit wonky).
• I volunteered to present at the September VMUG (Tarik, you too!)
• The next version of the Ugly Bugbook should be out shortly.

I'll keep chipping away at the projects and posting the news here.

Assuming too much?

Gah!  It's 3 a.m.  I've spent the last two hours troubleshooting an "adapted" function in Statusnet (yeah, I still use that).  Traced the problem to a file that doesn't even belong to Statusnet, isn't in the path for Statusnet, and shouldn't be detected by Statusnet.  It gets detected because it's called "settings.conf" and is higher in the file tree than Statusnet.  Once I changed the filename, and the two lines of code in the program that actually uses the file, everything in Statusnet went back to normal.

The real head-scratcher is that the file has been in that location (and in use) for months.  It should have caused problems last year, rather than waiting until the middle of May to cause problems.

I guess this one will fall into the "we'll never know" bucket as I'm not willing to pursue that minutiae that far into the rabbit hole.  However, I will say that the Statusnet programmer may have been a bit presumptuous when thinking that the only program on the web server (using "settings.*") was his.

Bleh.  I'm going back to sleep.

Teaching an old dog old tricks

Was picked up as a Road Tester for the Wolfson Audio Card for the Raspberry Pi.  As my project is a bit complicated (want to include a LCD display), I'm having to relearn skills that I haven't exercised in years.  Specifically: soldering.

Used on old prototyping board for the first attempt.  This turned out to be a good idea as I absolutely suck at it.  Need more practice!

How long does it take to wear out an SD card?

In answer to the above question (in my case), around 16 months.  I'd been running a collection of internal services (BIND, Bitlbee, Openfire, rsync, Hamachi, Apache, etc.) on a single Raspberry Pi.

The SD card held up for roughly 16 months before the system started acting wonky.  Wonky, as in some services would start, others wouldn't.  There was no rhyme or reason while troubleshooting.  Finally I gave up and tried an install, using a new SD card.  This appears to have fixed it as the system is operating normally again.

The pain caused by the above was minimal, but only because I had backups.  Recovery took only a couple evenings work.  I shudder to think about what I'd still be doing without those backups.

NFC Ring Project

I believe I've found my next project.  It appears that the NFC Ring is about to hit the streets.  There's a bowl on my dresser where I toss my keys, pen, and whatever else is in my pockets, when I arrive home.  I'm thinking that I could marry a Raspberry Pi, a NFC sensor, and a NFC Ring, so that the action of dropping the ring into the bowl (or other container) causes the lights in the office to come on and a WOL packet is sent to the computer.  Inversely, removing the ring from the bowl should ensure that the lights are off (after a certain period of time).

I got the above idea while researching NFC tags.  I'd seen a project where someone drops their phone into a bowl that has a NFC tag in the bottom, at his home, office, and gym.  Each of the tags causes the phone to change settings, join Wi-Fi, etc.

The ring project looks a bit more attractive in that I only need it at one location and the charge on the phone remains an issue (can barely make it through the day on one charge).  What'd'ya think?

- Tim

The 757 web site has been wobbly of late and users are starting to complain about the amount of space taken up by 11+ years of blog post and wiki entries.  As such, NeighborhoodTechie will become the home of the blog and the wiki is being converted to book form (more about this later).

PDF text extraction tools

Building a larger tool out of a collection of smaller tools can be quite a learning experience.  For the past few months, I've been working on a document search engine to hold and index a collection of PDF files which were generated via the PrintFriendly browser app.

In the past month, I learned a few things about PDFs and extracting text from them.  

- PDF is not a document language like DocBook or HTML.  Rather, it is more of a type setting language, in that letters are located individually on a page.

- There are no good tools to properly extract text from a PDF (commercial tools included).

- Most text extraction tools cannot properly handle the letters "f", "o", "ll", and "t".

Of the various tools tested, it appears that Calibre's ebook-convert produces the cleanest straight-text output.  I'm using that in the text extraction piece of the search and including the ability to edit the extracted text (to improve the searches).