April 1, 2008

Kevin Ludlow and I crossed paths around 2006, right when “myspace tracking” took the world by storm. Maybe Kevin’s tracker wasn’t the first, but as far as I know he was the first to come forward with a tracking system that honestly worked, intended to be used with Myspace. Now Kevin Ludlow has something new: openFace.

PJ: You’re not shy about sharing your life online. Do you publish everything or do you have some privacy controls for friends and family that request not to be published?

KL: This is actually a topic that has come up a little bit lately and one that is definitely not easy to answer. Firstly, I DO have some methods in place so that I can ‘privatize’ content, but it was not built to be used in the way one might think. The privatizing functionality is really just because I am trying not to make content public without it first being described, tagged, labeled, etc. That being said, there is actually a considerable amount of content on the site that is currently not visible to the user, but only because it has yet to be methodically (or programatically) categorized in some way.

Regarding my tell-all philosophy, it’s true, I am in no way shape or form bashful about sharing my life. Some people have noted that I might have embarrassing moments of my friends or family online, but for every one of those, there are at least ten of my own embarrassing moments. On the actual issue of privacy, I should start by saying that nothing I have has been captured without the person’s knowledge. Rare exceptions include things in public, but never breaking the so-called ‘reasonable expectation of public privacy’. Regarding actual censorship, I basically just have three personal restrictions:

1) I won’t write about or post any sort of media regarding any sort of sexual encounter. What might that include? I’ve been working on putting up my online journal for years now and though I might (and do) talk about girls I’ve dated, perhaps how we met, and etc., I’ll not disclose anything sexual about them. It’s really the only off-limits topic that I have as far as disclosing story details.

2) As a courtesy to my friends, I wouldn’t publicize anything criminal. Not that its ever happened before, but if I filmed a friend graffiting an underpass for example, I would not post it online - especially if the subject was visible or labeled. On the other hand, if I happened upon a random person committing a crime and filmed it, it would be public within the hour. Though there might be some hypocrisy to that, it’s just a personal, and I think somewhat obvious courtesy to my friends.

3) I won’t disclose anything that is ‘business confidential’. Simply meaning that I AM given privileged information from time to time (perhaps even just from friends) and wouldn’t want to break contractual or fiduciary responsibilities.

PJ: Do you collect any stats, like “most viewed album” or “top ten tag searches” and will you keep that info private or publish it somehow?

KL: I do keep stats on a few things. Nothing which is personally identifiable, but just so that I know what picture, videos, and/or texts are more interesting to people. I keep up with photo views, video views, and in the future will have any other mediums keeping up with basic stats as well.

Search queries are also kept, but not really so that I know WHAT is being searched per se, but rather HOW I can improve the WHAT people are looking for (if that makes sense). The system will report to me if someone searched for something that yielded zero results, or for that matter results under a certain threshold. If I were to find that people were consistently searching for something - perhaps ‘guitar’ - and zero results came up because I had a gallery labeled as ‘gibson’, it would prompt me to either create some sort of fuzzy search property between them, or to relabel my picture(s) to include the ‘guitar’ description.

PJ: Or take more pictures of guitars? Haha. Do you have any comment on the success of Photobucket? Photobucket seems like such a simple script, yet it’s phenomenally popular because the average social network user doesn’t have an FTP account.

KL: Although I haven’t really used Photobucket, I definitely can appreciate its presence (and other similar ones) on the web. I think it’s a fantastic tool for people that don’t have
a) access to large data hosting and/or
b) the knowledge to efficiently use data hosting.

I don’t know exactly how they [Photobucket] handle images, but the only concern / problem that I’ve ever expressed with online photo storage houses is loss of original data. Not losing the actual photos, but possibly removing important information FROM the images as to make them smaller on disk. For example, re-compressing the original photo format to x% so that the image becomes smaller on disk.

Again, that might be more of a thing from yesteryear - but it was certainly prevalent when I first set out designing openFace. My system is specifically designed to NEVER lose any of the key data that makes up the photos (quality, metadata, etc). Put simply, I always want the data in its most original state of existence. The basic theory being that you can always make nothing from something, but not the converse. If you re-compress an image and discard the original, you’ll never get those pixels back. …an extension of the basic bottleneck principle.

PJ: Yes, nice to have the original online, especially if you’re bad about backups, like me. The other problem with Photobucket, they run it like a business, which is why when your pictures pass some piddly bandwidth limit your picture transforms into an ad for Photobucket. Annoying! So these days people have hundreds or even thousands of photos. Just how comprehensive is your project?

KL: It might help if I provided a few of the detailed statistics regarding my current site. As of this writing, there are 18,800 personal photos visible and about 2,000 more not yet visible This gives a current total of about 43gb of data. While that number is honestly pretty small to me, it’s hugely unmanageable to most non-technical people. Even people that have similarly large mp3 collections (or the like) probably can’t appreciate how difficult it is to move that data-size around on the web without legitimate commercial network support.

If it’s still unimpressive, the videos might make it sound more interesting. I currently have 1,558 searchable videos (around 40 hours or so) and another 400-500 not yet visible (hidden for lack of descriptive content). The video data takes up closer to 900gb of data. Every video that I’ve ever shot (since I was a kid) is being digitized and added to my online collection. Not all 900gb is available online just yet, but it’s slowly being added.

If it still seems simple, keep in mind that all of the development is done on a single computer at home and I’m limited to transferring all of this from a cable modem. I’ve had processes running for literally months on end helping me to get the content online. Ideally I’d have all of this in a local data center, but I just haven’t the money to front for a personal and not-for-profit project like this. A personal website closing in on a terabyte of data is a hell of a lot of work for one person to keep up with.

As far as the future holds, when I finish the next part of my project which includes digitizing family films dating back to the 50s and having every print and slide photo in my entire family history digitized (dating back to the late 1800s), I’m guessing there will be closer to 30,000 photos online and perhaps closer to 80 hours of video). It’s an incredibly time consuming process for me. Given those and some other plans I have, I’m guessing I’ll be managing closer to 1.75 terabytes of data at that time (shooting for the end of this summer, but we’ll see how time permits).

PJ: Interesting. I used to work at a frame shop and people constantly came to me asking how they could archive old family photos. I got to ask a librarian about this, she wrote her masters thesis on plastic. As you can imagine there’s millions of old photos out there in plastic sleeves right now, but plastic hasn’t existed long enough for us to know how well it protects (or harms) our pictures, add to that most old photos aren’t printed on acid-free paper. Decay is a real problem and there’s no simple solution. Anyway. You’re obviously hosting these pictures yourself, do you use FTP? Could you see opening your site to other people, where family or friends could upload pictures without an FTP account?

KL: Although I could, I happen to not use FTP but rather transfer data over SSH/SFTP. At the moment I have special back-end scripts that synchronize my development computer (at home) with my remote server.

As far as extending functionality to other people, the short answer is yes. The archival system back-end is incredibly flexible and I have certainly considered how/when I will open it up for others to send data directly to the remote host. As far as the method of transfer would go, I’m actually working on a few ideas. There is always the simple method of using HTTP posts, but this seems impractical for large data sets (bearing in mind I really want to preserve content - not destroy it). A more likely candidate would be a cross-platform binary (executable) that would manage the process(es). I’m not particularly fond of Java so my guess is that I would write this future application in C#. It would be easy enough to compile it for Windows, OSX (my brand of choice), and even Linux (2nd brand of choice).

There has also been some consideration [from others] of wanting to extend the software for various other worldly applications. Given it’s just all talk at the moment, I won’t speculate on any of that at for now.

PJ: What do people like/dislike most about your website? What drives you to work on it, a passion for photography?

KL: Definitely the responses I’ve gotten. Most all of my friends have conceded that while they may have given me some hard times for my never-without-a-camera persona, being able to look back at years of memories is just overwhelmingly enjoyable to most of them. My girlfriend has been all too supportive of my work and I can’t thank her enough for that. I can’t say I’ve had a lot of development slumps, but she’s definitely reminded me of what my motivation is. Even friends and acquaintances I haven’t spoken with in years have written me expressing their gratitude for my love of preserving life’s memories and for including them in those memories.

If I could get on a brief tangent, the first real responses like this actually came from my parents, aunts, and uncles several years ago. The response came from some videos I had released of my grandparents shortly after they both died in 2005 (within a few months of each other). A few years prior to their deaths, I took 30 minutes out of a random January morning to interview them both about where they were born, their upbringing, things they had accomplished, and etc. I eventually digitized these videos, added transcripts to them (a very tedious process without tools), and tossed them on my previous website for public consumption. The response from the family was so enjoyable, and perhaps even a bit touching, that I felt the motivation to continue designing/building something that was really capable of disseminating these sorts of memories. Although I was already designing the system, this probably paved the real mindset for openFace.

PJ: If this was my website I’d spend countless hours editing the photos, cropping, color correcting. I’m a relentless editor. How much time do you spend editing your albums, browsing them?

KL: Well, that’s one of the big niches to openFace that I haven’t really discussed with too many people just yet - and in my opinion one of the better ones. Perhaps then this is a good medium for that question! There is definitely an ongoing time commitment that comes with keeping your life online, but the plan has been to make it no more time consuming than writing a simple journal / blog entry or anything else like that.

Although many of them are still in development, I’ve worked on/with countless AI technologies to make the system work faster and to be more hands off. Face detection and recognition algorithms, speech extrapolating techniques, image object recognition, just to name a few. There are even some systems that allow for other people to help me organize my personal data in fun sorts of ways. The processes that ARE in place seem to be working pretty well, and I’m confident that the other ones will move along in time. If I could develop the software full time, I would, I really would, but unfortunately I do keep a full time software development job too! There are just never enough hours in the day.