The Substantially Similar Weblog

I just had one of those gnucash days.

Every couple of weeks, I decide to sit down for an hour or two to catch up on my personal accounting. Suddenly, the sun has set, my legs have atrophied, and I realize I’ve forgotten to have lunch and dinner. Does this happen to anyone else?

Ostensibly, tracking personal finances should be a relatively simple matter for someone like myself without substantial assets or investments. But gnucash (or any other accounting program, I suppose) makes it difficult to “fudge” any numbers—if you started the week with $27.61 in your wallet and ended with $19.05, you need to say where that money went. I do have “unknown income” and “unknown expenses” categories, but they really bug me so I try to keep the totals down to a minimum.

The bigger problems are reconciling bank transfers, mortgages and loans (interest and principle), etc., when my financial institutions don’t always provide the best online tracking and reporting. For example, my bank has a “current balance” through web access, but that balance might be ahead of what the ledger shows (i.e., there are transactions reflected in “current balance” but not on the ledger)—and the “current balance” itself can be a day or two behind schedule.

I also wish gnucash were just a little bit better. It really feels like it’s been stuck for a couple of years now while all the other applications I use regularly (Firefox, OpenOffice.org, Gnome generally) have gone through major revision, and are genuinely fun to use.

In August 2003, the gnucash team issued a call for help. I admit I did not rise to the occasion, but I did expect more developers to respond. Gnucash seems to me to be an ideal open source project to attract hackers—almost everyone needs to use some kind of personal finance package eventually, and there are innumerable “scratch-an-itch” type clever solutions that could be implemented (automating web transactions with financial institutions, for example).

But Gnucash is brittle. The user interface is often awkward, and it’s awfully easy to lose work you’ve done. For example, clicking on “close” on a report makes the report just plain disappear, but clicking on “close” while looking at an account just closes that ledger. If you get into the custom of clicking “close,” you’ll lose the report you just created. And if you click “yes” when exiting you’ll lose all the work you’ve done in the session if you haven’t saved. I realize this is typically how word processors work, but I think many of us have different intuitive expectations from a finance program.

Here’s a solution to one problem I’ve encountered repeatedly, just to make this blog entry at least somewhat useful: if you move your account files to another computer or path, you’ll lose all your reports. They’re still there in ~/.gnucash/books, but they don’t go with the data files. If you move the data file *back*, you still won’t have your reports, because gnucash will have “forgotten” them.

If you look in ~/.gnucash/books, you’ll see a file corresponding to the path of each data set you have, with %2F substituted for /. Copy the file there to the corresponding name of your new data file.

Now edit ~/.gnome/GnuCash and find the section that corresponds to your old data file, and create a new section with all the same data except the title ( [MDI : ]). That should do it.

If someone knows a simpler way to do this, please let me know. Also if anyone has any inspiring ideas about what might move the gnucash project forward, I’d be interested in hearing ideas.

I’m not sure which possibility is more depressing: that Bush may have stolen the election again, or that he won without stealing the election. In either case, we’re in sorry shape. I’m not quite ready to join the “I’m moving to Canada” crowd, though—I don’t find that rhetorical tactic all that effective.

If nothing else, at least they won’t blame Ralph Nader this time.

Is anyone aware of an MP4/AAC sound file ID tag editor for Linux? There doesn’t appear to be anything in Debian; a Google search doesn’t reveal any obvious candidates.

Steve and I have decided to write and package freevite. Freevite will be a web-based event invitation/RSVP system licensed under the GPL. Before we start, though, I’m soliciting suggestions on the program. As far as I know, there is no free-as-in-speech web-based invitation package, and the world sorely needs one as commercial proprietary competitors brainstorm new ways to build a revenue stream from their product. It shouldn’t be that hard or time consuming to code, and it seems to me that we need a free-software standard product for this.

The program will probably be coded in perl, and give the administrator the choice of a simple file-system based data storage system or MySQL/PostgreSQL for better performance and data integrity. It should be possible for the administrator to plop the perl script down in a cgi-bin directory and set the proper permissions on the data storage directory and have a totally functional system, but also provide options for more secure (e.g., against cross-site scripting vulnerabilites) and sophisticated installations. The program will also be available as a Debian package, and support various ways of doing site-wide and user-specific installations (this will be tricky—Debian doesn’t seem to have a good solution for having packaged perl scripts like blosxom run on a per-user basis).

Administrators can configure the system so that anyone can create an event or only authorized users can create an event. There will be configurable privacy settings—if the user doesn’t want the inviter to know they’ve opened the invitation, they can indicate that, and the inviter will also have the choice of deciding whether to track opened invitations at all. You will be able to respond to an invitation without any registration or authentication process (having received the token for your invitation by email), but eventually there will probably also be a way to create a persistent identity linked with an arbitrary number of email addresses if you desire.

All presentation will be done with mailman-like templates, and the package will ship with some standard, clean, templates. Content will be properly separated into stylesheets and HTML.

Everything will be HTML standards compliant, and should render properly in text-based web browsers like w3m. There will be no required Javascript or any plugins, although there might be some optional Javascript content (my web-based photo gallery software, salonify, works this way).

Please leave suggestions, ideas, or recommendations as comments to this entry or email me. If there already exists a free-as-in-speech product that does all this, please let me know and I’ll stop right now.

In case you haven’t heard about it elsewhere, check out p2p-politics, a collaborative website where anyone can post their election-related homemade advertisements. I particularly recommend the “If the Bush Administration Was Your Roommate” series.

Now there must be some Bush supporters out there with an ounce of creativity and a digital video camera who can upload some content. It’s a little embarrassing to have such an imbalance there. Does anyone understand why there are hundreds of grassroots pro-Kerry ads, but not a single pro-Bush one?

Apologies to any Planet Debian readers for having just monopolized the front page with several stories; I moved a bunch of old stories to a new blog topic, and apparently Planet Debian thinks they are all now new stories, even though they have their old timestamp. If anyone knows a remedy for this, please let me know. (I guess this entry is further compounding my overpresence!)

My friend, fellow Boston resident and Princeton alum, Kerry-fundraiser extraordinaire, and non-blogger Jon Garfunkel·, suggests that the “mysterious lump” on Bush’s back in the first debate (as reported by Salon and later followed by the New York Times) is in fact a portable defibrillator·, just to be extra safe in these difficult times.

Jon has asked people to have spread the word to see if anyone can get their hands on one of these devices and find a 6-foot tall, 190 pound man, wearing a 44L jacket over it, and see how it looks.

Please let me know if you have any leads.

The instant debate polls are already in (well, they’ve been in for hours now). I thought this snippet from the USA Today poll was kind of funny:

B. Expressed himself more clearly 

Is it just me, or did the identity of Factcheck.com change overnight? I checked it when Cheney mentioned the site during the debate last night, and it was just a “portal” banner ad type site. Cheney actually intended to refer to factcheck.org, the Annenberg Center Political Fact Checking site (which is now giving a Microsoft .NET “Server error in ‘/’ Application” error—too bad they didn’t use Apache!). But when visiting factcheck.com today, it redirects to georgesoros.com, a message from billionaire and anti-Bush activist George Soros. Is it possible that he took over the site that quickly, or was the domain hijacked last night, or is there some other, more sinister explanation?

This debate is fairly nasty, and I can only see it getting more aggressive. I’m not going to “liveblog” the whole thing, as are many others.

What’s remarkable is how sharp both of them are. They are also flatly contradicting each other. The question is whether people will go and research the facts, or go with the candidate who they “trust” more at a gut level.

Interesting that Cheney referred to “factcheck.com” on the Halliburton issue. He actually meant factcheck.org (I’m sure the ad site .com got a few extra hundreds of thousands of hits from that, though). Although factcheck.org is down right now, probably also due to server load, a glance at the google cache of the page reveals many articles critical of the Bush campaign’s claims. I wonder if Cheney is just counting on people not actually checking the site, and just figuring it refutes the Halliburton attacks.