A few weeks ago in my EMT refresher class, the instructor mentioned something about the movie Airplane! to one of the students. The student, who was about 21 or 22, didn’t have a clue what the instructor was talking about. While I can’t remember if I saw the movie in theatre (probably not because I was 7 when it came out), I did see it and the sequel sometime during the 80’s. I really felt like I was getting old because it didn’t seem all that long ago that the movie came out.
Stop wrapping dmg files in zip archives!
I routinely download software that ships as a Mac disc image, but when uploaded to a web server is placed in a zip archive. The dmg file is already compressed, so the zip file doesn’t reduce the size that much and just adds one step to decompressing (when I decompress, I’m left with the zip file and dmg file; I then have to trash the zip file). My suspicion on why this is done is that the web server isn’t setup to send dmg files as binary and people get a mess in their browsers. So solution (for apache servers), is simple. In /etc/mime-types, add dmg to the octet-stream line like:
application/octet-stream bin dms lha lzh exe class so dll img iso dmg
First look at OCR
A recent comment on my blog struck a nerve where the commenter said that OCR would basically put a competitor to ReceiptWallet above it. While I still don’t believe that OCR is all that useful for receipts (if there is one mistake when you’re generally only entering 3 small pieces of data, you’ve wasted time because you have to review each entry carefully), I took a look at an open source OCR package. While this code is a bit rusty, there has been some recent work on it. My first test was a Rite-Aid receipt where I was looking to see if it could read 3 pieces of data, the merchant name, date, and total. It failed on the merchant name because it was a graphic, however, it picked up the date and total in such a way that I could parse the data and grab what I needed. I then tried 2 other receipts, both from Costco and the results were completely miserable such that I couldn’t get anything from them. I’ll keep plugging away and testing to see if my results are better.
In addition, I put in a request for a quote for a commercial OCR engine. However, I suspect that it will be cost prohibitive. If it costs $5,000-$10,000 upfront plus a per copy licensing fee, I can’t afford that as it would completely wipe out any profit unless I significantly increased the cost of ReceiptWallet.
If anyone has more information on OCR engines for the Mac (commercial or open source), please let me know.
Goldilocks and the 3 keyboards
I somehow managed to spill stuff on my keyboard (not the internal one in my MacBook Pro) and have been looking for a replacement for awhile. I dried it out enough so that it has mostly worked, but I kept missing keys. So I bought an Apple keyboard from CompUSA a few months back (when they still had a San Diego store) and had 2 problems with it; first off the unit was defective as not all the keys worked and second, it was too small for the keyboard tray that I have (it has spring loaded holders on the side). I returned it and ignored the problem for awhile. On Friday I went to my most hated store, Fry’s, and got another keyboard. It looked OK, but when I got it home, I discovered that the command key was too small. My choice of wired keyboards is quite small for the Mac; I want wired so that I don’t need drivers and it works with my KVM switch. I also want a Mac keyboard as the Windows keyboards that are Mac compatible have the command and option keys switched unless you install the drivers.
Back to Fry’s today and I picked up a macally keyboard. It seems to work well, doesn’t require drivers (they have drivers, but I have no idea what they’re for) and fits my keyboard tray. The only problem is that I paid $10 over retail. It’s not worth my time to return it to Fry’s and then order it online (or maybe it is as it is $45 from Amazon with Free shipping).
Running my own server vs hosting elsewhere
With all my problems with my server, you’d think I’d give up and just host my server somewhere else and make it someone else’s problem. After a little thinking, I’ve come up with the pros and cons of running my own server.Pros————–Physical security of serverMusic server – I need a server to run my SqueezeboxesUnlimited bandwidth (at 1 MBps upstream, but unlimited)Full backups – I backup the server to an external drive and take it offsiteLots of storage (currently at 300 GB, but only about 15% used)It’s a learning experience5 static IP addressesFull control over serverAbility to run my own PBX and have good voice qualityCons————Electricity usageI have to keep it runningLimited upstream bandwidthPotentially more costly (it costs about $50 per month for me to host my own server); however for the hardware I have, a hosting service might charge more for RAID 1 and backupsWhile there are more items I’m missing here, it’s pretty clear to me that running my own server is the right way to go. My issues last week appeared to stem from drives that were failing; it just happened that when I was putting in the new drives, the drive I left running was failing causing some issues. I’m now up and running on my 2 new drives and have been for about 5 days now. Lesson for the future…replace drives about once a year if they need it or not as the drives are relatively cheap ($150 for both drives) and run 24/7 which puts a lot of wear and tear on the drives. Now the question is, what do I do with the old drives? Would a reformat on the drives do the trick and keep them running? What less mission critical role can I put them in?
The new world of IMAP
I realize that IMAP isn’t new, but I just started using it yesterday. I’ve resisted for years as I’ve always had one main store for my email. However, lately, I’m checking email using web based email and everything I’ve heard about IMAP make it sound like it is perfect for using in multiple locations. So far it’s working OK, but I’m getting used to it. For instance, if Mail on my Mac filters messages out of the Inbox, they’re pulled off the IMAP server which kind of defeats the purpose of using IMAP. Most of my email is filtered. I created a rule in Mail that moves all messages < 1 day old into my Inbox and made it my first filter. This should leave everything in my inbox, but when I run the filters over my inbox manually later, it should put everything in the right spot.
Hopefully I’ll get used to this soon; I’ve been using POP3 for close to 15 years ever since I started using Eudora in college. It’s hard to change my thinking, but since I started using POP3 over SSL, I could no longer telnet into my server and issue commands (I know the POP3 commands as I’ve written 2 POP3 email clients in my life), there really is no advantage to me using POP3.
I’m alive!
Yes, it’s true, I do have a heartbeat and this EKG proves it:
Useless application?
I downloaded and installed Little Snitch today to give it a spin as it was part of the bundle of software I bought it (which brings me to the question, why did I buy it?, but that’s another story). The program is well implemented, but just about every application makes outgoing network connections these days, so it is always popping up basically saying “xyz application wants to connect to abc server. Allow?” After click always allow more than a few dozen times, I finally disabled it (for now). I’m sure after I train it, it will be less annoying, but how can I really tell what applications are making legitimate requests and which are not? This seems like the little boy who cried wolf.
Speeding up Ruby
When I setup my store, I sued FastCGI to get acceptable performance which was fine as I don’t get many hits. Today I installed BrowseBack because it came as part of the software bundle I purchase through MacHeist/MacUpdate. As I hit my store a lot to check stats and such, BrowseBack kept loading the store which spiked the server load to 4 and required me to kill BrowseBack to get my server back to normal. While this isn’t normally a problem, it exposed a potential kink in my server’s armor. So I went looking for an alternative. I’ve read about using lighttpd by proxying requests from Apache. Setting it up was straightforward and seemed to work well, until a customer complained that he couldn’t purchase. I tried it myself (again) and it worked fine. I had someone else try again and it failed. Hmmm. After much tinkering, I figured out the problem, my store code required an https connection otherwise it redirected to https which was fine, except that coming from the proxy, it was always an http connection. Since I already do a redirect on my store so that any http://store.receiptwallet.com requests goes to https://store.receiptwallet.com, I can be assured that all requests are secure. So I commented that line out of my store and everything is working fine.
(On a side note, the reason I couldn’t see it myself is that it was checking for local requests and since I access my server using a private IP address range, it basically got flagged as local, so it never required the SSL connection.)
Am I being told something?
I went to install a new hard drive in my server today to pre-emptively avoid a disaster as the drives in my server have been running 24/7 for 1.5 years and the drives are so cheap, that replacing it should save me time later. I thought all I had to do was power down, yank one drive, power up, partition the new drive, and rebuild the RAID. That is great in theory, but my RAID had errors so when I rebooted, it failed to reboot. I ran fsck on it (after failing to read what the screen told me to do for too long) and let it repair lots of little problems. After that, I was able to reboot and rebuild the RAID. In about an hour, my RAID will be rebuilt and I’ll let it sit for a few days before I go ahead and replace the second drive. With all the problems I have with my server, you’d think I’d learn to just use a hosting service and let someone else manage the hardware. Oh well.