Last week Ivan Misner asked me to add some features to the BNI Podcast website, namely the Digg Digg  and the Translate This Button plugins. Installing plugins is a simple enough task, as anyone who works with WordPress knows. Configuring them, however, can take some time. I determined that a horizontal layout was best for the Digg Digg sharing buttons given the fact that they showed up right under the podcast player. The translator is normally meant to live in a widget; to put it elsewhere, one needs to edit the theme files. Again, this was a fairly painless process.

But it was getting late and I was getting sloppy, and I didn’t do what I usually do, which was to make a copy of the unedited files before changing them. In the case of the page templates for the blog index, single post, and “static” pages, that didn’t matter: I got it right the first time and what I did was simple enough that I remembered all the steps and could have undone it. I still don’t recommend that, however, because after doing all of that successfully, I overstepped myself.

I was just checking that everything I’d done looked right on the site when I noticed that the photo of Dr. Misner on the “About” page wasn’t aligned correctly. And, in fact, when I tried to align the photo to the left of the text, it worked in the visual editor, but didn’t display properly on the public-facing page. “Hmm,” I thought, “I wonder whether Brady forgot to code the image alignments into the theme.”

So I went and took a look at the stylesheet. This is a text file full of CSS code that tells browsers where to put images and design elements and what colors and fonts to use when they render web pages.

And here’s where I made my big mistake.

A few months ago BNI launched a huge re-branding effort and there was a minor site re-design with a new logo, header and menu style. I wasn’t the one to do that work; the copy of the stylesheet I had on my computer was actually from the previous redesign.

And I forgot to look at the dates on the files and download the new one to work on before adding the code for image alignments  and captions. That’s an embarrassing n00b mistake. I should actually have downloaded a full copy of the theme files as soon as the new design was up there, but I didn’t. I can’t even say that I thought I’d done it. I didn’t think about it at all.

So I added the new CSS code to the old file, uploaded it to the server, and overwrote the stylesheet that told WordPress where to find the new header and how to display the new menu.

Oops.

I would have been in very serious trouble if Brady hadn’t had a local copy of the CSS file that I had just obliterated. This time I saved a backup copy of it with a different name before adding the code for the images.

Here’s what I should have done, and indeed what I normally do:

1. Downloaded the complete theme files that were on the server and saved them in a folder called <theme name>-<last edit date>.
2. Compared that date with files I already had on my computer, and also, if necessary, compared the contents of those files. Renamed my local copy of the theme to <theme name>-<last edit date>
3. Created a new set of theme files to edit, either by downloading them again, or by copying them locally. (Most WP theme files aren’t large and don’t take long to download, unless they have both a lot of custom page templates and a lot of images. This one doesn’t.)
4. After the work was finished and the files uploaded, renamed the working folder to <theme name>-<current date>.
5. Backed all those files up in my next set of local backups.

This is a manual method of doing what software developers know as version control or revision control. Real programmers use version control software like Subversion to handle this kind of thing for them, particularly if they’re working in groups on the same project.

That seems to me like overkill for websites, where the code tends to change only once in a while rather than undergoing continuous development. But it would probably be a good idea for me to start backing up the site’s theme files on a more regular basis. (The content of the BNI Podcast site changes much more frequently; I get weekly backups of the database that stores that e-mailed to me, and I already have local backups of the audio files.)

Whether you’re dealing  with CSS or Microsoft Word, it never hurts to save a copy of the original document before you start messing around with it. There are times when the “Undo” function can’t help you.

At the beginning of June, I got a new computer. It was high time: Enna is more than four years old now, and she’s pretty sluggish, and Windows XP is getting just a tiny bit outdated, though I preferred it to Vista.

The new machine is a thing of beauty: a 17.3” ASUS G72GX with 6 GB RAM and a bunch of other drool-inspiring specs, plus flash gamer details like a backlit keyboard (the Ur-Guru especially liked that one) and a glowing blue Republic of Games logo on the back. Seen in the sunlight (a rare occurrence), the paint reflects blue sparkles.

And, of course, it comes with a whole new operating system: Windows 7. So far I quite like it, but it’s a big jump from XP, and I’ve just checked Windows 7 Annoyances: Tips, Secrets, and Solutions (affiliate link) out of the library to help me find my way around.

The new OS isn’t the only reason to re-evaluate my existing backup system. Enna had two physical 80 GB drives, C and D. I used the C drive for both programs and data/documents, and backed up data and documents from the C drive to the D drive using SyncBack Freeware. I also backed the C drive up to both a USB drive (Ruby) using Karen’s Replicator and to a NAS drive (Teratides) using the built-in Maxtor backup software. I backed the D drive up to my Buffalo Quattro (Qualora) using SyncBack, and likewise backed Ruby up to my second NAS drive, Lachesis. (Confused yet? I posted a map in February 2009, when I was still using Freya instead of Ruby.)

The new machine, Auset (that’s how scholars think Egyptians pronounced the name of the goddess we know as Isis), has one drive of 500 GB. It’s partitioned into a recovery area (50 GB, no drive letter) , the C drive (116 GB) for the operating system and program files, and the D drive (334 GB) for my data.

There would be no point backing up from my C drive to my D drive on Auset, because they’re actually the same hardware. Backups all have to go somewhere else. On the other hand, the new folder structure has made my Replicator Jobs List very short. Instead of a dozen separate jobs backing up different folders, I only have to back up the “data” folder and the “docs” folder to appropriate places on Ruby, and all the subfolders will take care of themselves.

I’m going to have to add at least one more job, I think, unless I can persuade Microsoft that it really doesn’t want to keep any documents on the C drive. (I did figure out how to set the “Docs” folder on D:\ as the default location in the library, even though I’m not really sure what this library business is about.)

I won’t say any more about Replicator here because I’ve written about this handy free tool ad infinitum, starting in 2004.

I also decided to check out the new Windows 7 Backup and Restore program. It’s been a long time since I looked at Windows Backup—probably not since my first days with XP. It left rather a lot to be desired at that point. I figured it was worth checking out, even though the Ur-Guru had already set me up with Acronis Backup & Recovery 10 for disk images. (I’ll talk about that in detail in a future post.)

You can find Windows Backup and Restore in the control panel—at least, you can if you switch to the icon view. If you view the control panel by category, you may be hunting around for a while.

As you can see (at least if you click through to the full-size version of this image), when I went  in to inspect, I found that my last backup hadn’t completed. This was probably because I’d elected to make a system image but not turned off all my programs when running the backup. (Oops.) I changed my settings to leave out the system image (it’s that checkbox down at the bottom.)

Then I shut down all my programs anyway, just to be safe. Some things, like your Outlook PST file and your Quicken data, can’t be backed up while the program is running. You can schedule automatic backups, but if they’re going to work, you have to remember to have those programs shut down at that time.

This time the backup completed successfully, and in a fairly short period of time. If I want to restore only some of the files, I can click the “browse for files” button after clicking the “restore my files” button.

Note: the “Recover system settings” op
tion sends you to System Restore.

(Notice that it does versioning, since it offers you multiple dates to restore the files to.)

Once you pick a file to restore, you get a choice of destinations:

And then (if you’re just testing, and didn’t actually lose the file) you get the Win 7 file overwrite dialog:

What the heck, at least it’s prettier than the Windows XP version of the same message, as well as having one more option.

As backup programs go, this one has come a long way since the first time I looked at the built-in Windows Backup tool. I haven’t tested the system image against other imaging tools (perhaps I’ll include that in the Acronis writeup next issue), but as file backup programs go, it’s perfectly adequate. It doesn’t have the flexibility of a SyncBack, where I can create several jobs and schedule them to run on system idle, but an initial examination suggests it’s the equal of several products I’ve covered here.

And just to make me glad I had backups, less than a month after I’d bought her, Auset experienced the Black Screen of Death. When I turned her on, nothing happened. No drive light. No drive sounds. Nothing on the screen. I mean, nothing. Naturally I panicked and called the Ur-Guru, while digging out the info I’d need for a warranty return. (We—he—found the solution, which was to unplug the machine, remove the battery, and press the power button for 45 seconds.) Throughout the whole harrowing scenario, there was one comforting thought: at least I had backups.

Last week I got a call from someone whose name I won’t reveal but whose story should be a lesson to everyone reading this. The person in question is someone who uses her laptop continually, but doesn’t know that much about it.

“My computer went black,” she said.

“Totally black?” I asked.

“No, it has white letters on it. It says ‘Operating system not found.’”

“Oh,” I said intelligently. “That’s not good. Have you been using the backup drive?” (I knew this person had a supposedly foolproof, plug-it-in-and-it-backs-you-up Rebit SaveMe drive.)

“No—it doesn’t fit on the desk.”

Beat. Head. On. Wall.

As with exercise, the best backup system is the one you’re going to use. This one had seemed just about foolproof, but you do have to actually connect the drive to the computer. And while online backup might be more foolproof, particularly for someone who’s usually connected, it’s not that practical for RAW photos in the thousands.

A simple reboot actually brought the operating system back, but only temporarily. The machine continued to crash, and the “check hardware” message made my client suspect (incorrectly) that the backup drive might be the cause of the problem, so she disconnected it before it could complete its backup cycle. (Though there really did seem to be something wrong there, as transfer speeds to that drive were insanely slow when I later tried to copy some photos to it directly. I had, however, tested the drive myself, and it performed normally with my own older and slower machine.)

A few days later, the Ur-Guru and I actually got to see the machine bluescreen. By this time I’d already looked through the Event Log without finding any explanations for the problems and checked the health of the hard drive with the Computer Management tool. We’d even cleared off about 100 GB of old photos so the drive wasn’t red-lining in the space department. Everything seemed normal, but obviously wasn’t—especially when the machine crashed again. This time it reported memory errors, so we tried running the Vista memory diagnostics tool. (I never use Vista, so didn’t know about this.)

The test reported no problems, but the machine was obviously having problems. The Ur-Guru downloaded Memtest86 to give the RAM a further workout.

Still no problems found. But the machine kept acting up. I hate problems I can’t understand, especially serious ones like that. At this point, voodoo starts to seem like a viable option.

The only remaining thing to try to see whether the problem was with the hardware or the software was to reinstall the machine—preferably with Windows 7 instead of Vista. But that would reformat the entire drive, erasing all the data. And there was no guarantee the drive (or other hardware) was really stable enough for that to work.

Or, of course, you could just replace the computer, in which case it might be possible to salvage the data by taking the drive out of the machine, putting it into a 2.5” drive case, and connecting it to the new machine by USB. But that’s a somewhat iffy proposition with a drive whose soundness is in question, and not something a novice computer user can do anyway.

This mysteriously failing laptop is less than two years old. There’s no real reason (well, apart from the Vista operating system) to expect it to have serious problems so soon. But you never know what’s going to happen to your computer, and that’s especially true for laptops, which are much more at risk of being dropped, spilled on, or stolen. That’s why you need to start backing your machine up as soon as you get it—or at least as soon as you start putting data on it.

In this case, it looks as if installing Windows 7 (and reformatting the hard drive) has restored the laptop to a functioning state, though it’s too soon to be sure. But everything that was stored on that 320 GB hard drive is gone now.

Don’t let it happen to you.

Just occasionally, I think I might have been writing this blog for too long. Only very occasionally, because it won’t be “too long” until everyone backs up, the technology to do so is completely mature, and no one needs to hear about the subject any more.

But the end of one year, and the beginning of another, is one of those times. How many ways are there to say “Okay, make a copy of all your project-related and financial data from the past year, copy it onto a couple of CDs or DVDs, and store it with your tax information”?

Well, let’s see.

I first addressed this subject in December 2004: “Year-End Backups.”

Why do you need a year-end backup? Primarily for tax purposes. Just in case you’re faced with the dreaded audit, you want all your financial information readily available. (The more organized you are, the faster the IRS will go away.)

In December 2005, the post was “Don’t Use CD-RWs for Year-End Backups.”

Strictly speaking, year-end backups aren’t really backups; they’re archives. You make copies of all your important computer files from the year in question to store with your paper files. You need to keep anything relevant to your taxes (like your Quicken or QuickBooks data, bank statements, invoices from vendors, invoices to clients, and so on) for seven years.

And in December 2006: “Is It Time for Year-End Backups Again?”

Most of what you need to archive at the end of the year is confidential or at least private. That makes it a good idea to password-protect any files or folders you are backing up. Outlook, Quicken, and QuickBooks have this function built-in. For other files and folders, you might want to use a compression tool like WinRAR which allows you to put a password on the archive file. You’ll also be able to fit more data on one disk this way.

Instead of a December 2007 article, I wrote the next post on this subject in January 2008: “It’s Time for the Annual Archive.” (I’d finally given up calling it a backup, since it isn’t, really.)

I’ve just made 4 DVDs to add to the tax box. […] The most time-consuming thing about making them was isolating 2007 data. In some cases I had already done this, but I haven’t been completely consistent. Once it was done, I removed all finished projects from 2007 from my C drive to make room for 2008 projects. I’m not that pressed for storage space on my machine, but it’s annoying to have to look through folders for clients I’m finished with, or previous versions of documents I’m working on, when I want to get to my current work. So I use making the annual archive as an opportunity to tidy up my hard drive.

Last year I appear to have skipped the annual archive article, though I dutifully went through and created my annual archives. They got as far as my Buffalo Quattro drive, and somehow never made it onto CD.

So here we are at the beginning of 2010, and it’s time to clear 2009 out of our computers. (Please!)

If you’ve been making annual archives all along, and only keeping your current projects on your C drive, then creating your 2009 archive should be a comparatively easy process. Just copy all your data onto appropriate storage media. What kind of media depends on the type of work you do. If you work mostly with text-based documents and spreadsheets, it shouldn’t be too difficult to fit your project files onto a few CDs or DVDs.

On the other hand, if you’re a photographer, create a lot of video, develop software, record uncompressed audio, or otherwise generate data in simply stupendous quantities, then you’ll probably need an external hard drive or some Blu-ray discs. (The hard drive might be cheaper, honestly, but keep it in an airtight plastic box.)

And what if you haven’t been making annual archives all along, and your computer is full of years’ worth of undifferentiated data?

Well, don’t you have some fun in store for you?

Fortunately, it’s not as bad as all that, because you can use your computer’s search feature to find you all the files that were last changed in 2009. And then in 2008. And in 2007. And so on. Then you copy those files onto your storage media for archiving. Once you’re sure that the archive copies work (insert the CD or connect the XHD and try opening a few files), you can delete the old files from your computer’s hard drive.

Wow! Look at all that space you have for your 2010 files.

Do the same thing for your e-mail, making sure to create a new archive folder for Outlook. (You need to re-name the existing archive.pst folder to something like archive 2009.pst.)

Of course, if there are projects from 2009 that you’re still working on, you should continue to keep them on your computer. But if you’ve closed off the job and aren’t working for the client anymore, the files can go to a separate location.

Though you still might want to keep more than one copy. The annual archive is just something to keep with your tax documentation in case you (or the IRS) need to refer to it later. It’s not really a backup. So having a second copy of those 2009 files could be a good idea. You never know when a client from four years ago will call and say “Do you still have that…?”

After last week’s meditation on using external hard drives instead of DVDs to back up photos, Loyal Reader Mike Van Horn responded with the following e-mail:

My advice to you is, “Purge!”

I look at my own backups of photos, and what I see is lots of garbage (at 1 to 2 meg each). Many duplicates of the same photo. Gobs of photos of people or things I never need to view again. Photos of all the notes I’ve ever taken, including notes from people who have sold their businesses—or died.

If I would go through my photos and select only those that I could conceivably want to view again during my lifetime (or that my posthumous biographers would draw on!), I would reduce my photo storage by 90%.

It does take some time to do this. But if you don’t, you’re backing up haystacks just because they may contain a needle. And in that case, the needle is already essentially lost.

Mike makes a very good point. I went through my old printed photos a few years ago, and was appalled to find out how many prints I’d saved that were out of focus, overexposed, grainy, badly composed, or whatever. In fact, there was no real reason for me to have held onto any of the shots that weren’t good enough to make it into the photo albums in the first place, apart from reflexive pack-rat-ism. And I had two copies of a lot of them, from the days when many discount film developers offered free double prints. Twice the trash! Out they went.

If you mess up a photo and you know it, you should delete it right away, before it even gets to your computer. Make room on your card for a better one. There’s certainly no point preserving junk in quadruplicate.

And with the end of the year coming up, it’s a good time to go through all your files anyway, to see what you need to keep and what you can pitch. Cheap storage shouldn’t lead you into sloppy habits.

In some cases, though, “Just throw it away” isn’t an option. Even small companies may be bound by data retention laws, depending on the kind of business you’re in. In that case, the best you can do is move the data off your computer into your archives, and practice what’s known as deduplication. The larger your company, the more space deduplication will save you, particularly when it comes to things like e-mail attachments that get sent to all 10,000 of your employees. You don’t need 10,000 copies of that file, only one.

Enterprise backup and archive systems all handle deduplication automatically. Some of the SOHO backup services are starting to offer it as well, the better to keep within your cloud storage limits. In most cases, however, the best judge of whether two files are identical (in spite of having different names or dates) is going to be you.

And no matter what system you use for storing photos—or anything else—make sure you organize and label them to better help you find what you’re looking for a few years from now when you’ve forgotten everything you think you’re going to remember about them.