OddThinking

A blog for odd things and odd thoughts.

Online Photo DB: Stage 1 – Problem Statement, Loose Requirements

This post is part of the Online Photo Database project documentation. Learn more about the project’s current status.

As a long-term project, I wish to re-host my photographs. I believe you do too, but perhaps you haven’t realised it yet. Therefore, I am sharing my thoughts.

The underlying problem is that I want to be make the photographs I take easily available to the people who are in the photographs.

I currently use four techniques:

  • Old-School: Photo albums and boxes of stray photographs and slides.
  • Limbo: Files sitting on my hard-drive and backed up on DVD.
  • Custom Software: I wrote some custom web software (starting with a burst of development in 2002, slowing down to a trickle by 2005). It hosts around 4000 photos.
  • Facebook: Since around November 2007, when an event is organised on Facebook, I have put up the photos of it there. That accounts for about 400 photos.

None of these techniques meet my requirements.

Which raises the question: what are my requirements? Are they really that special?

I am glad you asked. I don’t think they are particularly special to me; I think you have most of the same requirements – and possibly some more. Let’s have a look.

Photo-database Requirements

  • The images shall be available on the web, accessible by a typical web-browser, without requiring registration for the typical user.
  • Print-quality (e.g. 0.5-1.0 MB sized) images shall be available.
  • Desirable: Archive-quality (e.g. 6-10MB) images shall be stored.
  • Web-quality (e.g. 30-200 KB) images shall be displayed by default.
  • It should automatically generated thumbnail and web-quality images based on the originals.
  • Desirable: Customisation of thumbnails and web-quality images shall be available (e.g. choosing to crop over shrinking.)
  • Critical: I own all images I upload. Me! Not you! Me!
  • I own all meta-data that is placed on an image. That’s right. Let’s be clear about this. Me!
  • All images and meta-data is either backed up to, synched with or sourced from my hard-drive that I can hold in my hands, read with my own software, and put under my pillow at night.
  • Don’t waste my time with quotas… Okay, I will accept a quota that is an order of magnitude greater than I will ever use. I currently use 2-3 GB online and 40-50 GB on my hard-drive, so don’t bother me any quota less than 0.5 TB.
  • Let me group photographs by event.
  • Desirable: Let me group events by type. (e.g. Let the mibologists view all of the Mibology Conventions.)
  • I need the ability to tag who is in the photograph.
  • I also want to tag places, pets, things and activities.
  • It was a big insight after a few years of tagging photos to realise I want to tag performers by their real names and by their stage names too.
  • Desirable: Being able to tag particular parts of the photo.
  • I would like the ability for other people to tag my photos.
  • Sometimes, I will only know the person’s first name. Sometimes, I will know their name, email address, web-site, Facebook ID, etc.
  • There will be multiple different people with the same name; let me add secret comments to remind me which one is which.
  • Their email address should never be published on the web, but a link to their web-site should be!
  • Sometimes, I will correct a person’s detail. It should update everywhere.
  • A key reason for tagging is so people can find other photos of the same person.
  • Even in different albums on my site.
  • Desirable: Even on different web-sites, such as Flickr or Facebook.
  • Some photos are not suitable for wide distribution. Have some decent concept of privacy. Think about Google when you come up with that.
  • I want to tell people when photos of them become available. No-one has ever complained to me that such an email is spam.
  • Comments can be fun
  • Comments only work if you can subscribe to comments; especially if you are in the photo!
  • Sometimes, I am not the photographer; let me assign credit.
  • Sometimes, the photographer is not the copyright owner. Let me assign ownership.
  • I often take lots (way lots) of photographs. Big collections of photographs can be boring and scare people off. The perceived quality of the album is (I believe) proportional to the average quality of the photographs, not the size. Thus, adding photographs to an album may actually reduce its overall value. Therefore, I want to be able to rate the quality of the photographs, and only present the best ones to the casual user. Interested parties should be able to lower the quality threshold to see more.
  • Read my EXIF data; particularly date and time info.
  • Desirable: Some control over the ordering. Particularly, I do not want to be forced to have photographs sorted by time-stamp.
  • Support search of tags
  • Support search of album names
  • I need to be able to find an original photo on my hard-drive based on an edited on-line photograph. Store some meta-data to help me do that.
  • Help me out with the tags; I shouldn’t need to re-type the whole tag every time.
  • Be wary that I have over a thousand different tags. You need to scale!
  • Let me keep images in a draft state until I am ready to publish them.
  • Desirable: RSS and ATOM feeds, for the win!
  • Next and Previous buttons when looking at a large image.
  • Desirable: Slideshows
  • Desirable: Don’t be scared to put more than 5 thumbnails on a page.
  • Fast viewing is good. Super-Duper fast viewing is perfect.
  • Cheap is good. Free is perfect. The biggest issue with “cheap” is that one day it might become expensive.
  • I already have a decent web-host. The incremental financial cost of hosting another application is pretty close to zero.
  • I take many photos, which makes optimising the administration work-flow important. About 1 minute face-time per photograph would be desirable.
  • I want to see constant improvement to the software, while I sleep.
  • Mandatory: Ability to migrate from my old software to the new system: That means an API of some description. That also means having tags that are not limited to some part of the photograph.
  • My current software permits me to construct a URL like so: example.com/allPhotosOf/KeanuReeves. That way, if I am ever at a party with Keanu Reeves, I can tell him the URL to see all the photos of him, without needing access to a computer. That’s nice (and relies on a search and non-clashing subject names!). It is desirable to be able to keep doing similar things.
  • If you are going to keep viewing metrics, the metric I want to optimise is “How long does a visitor stay?” so be sure to tell me that.
  • Anti-requirement: You might need the ability to do simple edits to the photos online. I don’t; I use Photoshop.

That will do for now; I will probably come back and update these with some more requirements later.

Next step is to look at my bespoke software, Flickr, FaceBook, Coppermine, Picasa Web, etc. (any other suggestions?). I will see which of the requirements they fail to meet.


Comments

  1. Note to self: Must improve appearance of HTML lists on OddThinking; there needs to be more spacing between items. I wonder which CSS item controls that.

  2. I can’t help thinking you should have numbered these ahead of time…

    Also, is this a set of requirements just for you, or is it also for people (sort of) like you? If it’s for people like you, you might want to spend some time defining what sort of photographer that is. If I had to guess, I’d say the defining characteristics here are:
    a) a moderate to high volume photographer, but not professional. You do this for fun!
    b) better than consumer grade equipment, but not premium grade. You care, but seriously, who has that kind of money?
    c) mainly landscape and social photos, rarely portraits. Who wants stilted, lifeless photos?

    These would guide what sort of features on the equipment the system can rely on, and what sort of skills the photographer has in the cropping and retouching dept. It also means the system doesn’t need to support portfolio building and maintenance, which could otherwise be a major drawcard for a (semi-/wannabe-) pro.

  3. A good list Julian, I’ll be interested to see what comes up.

    FWIW, I currently use a combination of iPhoto, with exporters for both Facebook and Flickr. iPhoto is the source of all meta-data (mainly keywords) and it is included in the export when sending to Facebook and Flickr.

    This is of course non-ideal for lots of reasons which I won’t even go into, and I’m in the process of migrating to Aperture and Flickr Pro. However if you view a local photo app/library as the canonical source, and the web interface as merely a projection of that library, does that relax some of your requirements for the latter?

    I’m not sure whether you’re committed to photo library management software, but some of the recent products (eg Apple’s Aperture, Adobe’s Lightroom) are very capable. For example, the sorting/tagging business can be very time consuming, but with iPhoto I can typically average <5 sec/photo for tagging and rating. I tend to take photos of the same people though so YMMV, but you should be able to get much less than 1 min/photo.

  4. Richard,

    I can’t help thinking you should have numbered these ahead of time…

    I considered it, but decided a short identifier could be bolted on later… along with some grouping by category… and clearer priorities…. and an IBM RequisitePro schema….

    As for the defining characteristics:

    Yes, I am a moderate-to-high-volume photographer, by old-fashioned film standards. Compared to kids these days with their new-fangled phones, I probably don’t count as such any more. I think the best way to describe it is that I am a binge photographer! I’m not sure if that affects my requirements much except to focus a little on the admin workflow efficiency.

    Yes, I have “prosumer” equipment, by old-fashioned standards. My aging digital SLR has less megapixels than those damn kids and their new phones. Again, I am not sure how this affects my requirements, except that I don’t need red-eye reduction filter built into the web-site.

    We may have different definitions of portrait, but I would have said that portraits account for a majority of my photos: parties and performers. I guess it makes tagging and commenting more important than for someone who photographs thunderstorms, but apart from that?

    I’m not trying to make money from the photos (although, I would probably consider throwing a few demure Google ads in there if it was an option.)

  5. Alastair,

    However if you view a local photo app/library as the canonical source, and the web interface as merely a projection of that library, does that relax some of your requirements for the latter?

    I certainly will be considering such models; this model meets the “I own the data” requirements well. I half expect that this is the way I will end up going.

    However, it doesn’t easily meet the “let other people tag/comment” requirements. I often don’t know the names of all the people I photograph.

    That’s not to say that it isn’t feasible; I just haven’t seen it yet.

    I am not expecting to find the perfect solution, so I will be weighing up the pros and cons of the various solutions available to me now. I only wish the current tools had been available in 2002; I never would have bothered to roll-my-own in that case.

    I have wanted to do a trial of Adobe Lightroom for a while; I thought I would find out what the new target of the workflow was before I assessed its ability to make it more efficient.

  6. Note to self: Here are some more demands to fold in:

    • Basic support for video would be desirable.
    • Within the definition of an event, you should be able to link across to other photo databases of the same event.

    There is another aspect that is harder to explain, and I need to think through more.

    The use of FaceBook photo albums introduced an interesting social networking aspect that caught me by surprise. Tagging of photos in FaceBook is limited to friends of the subject. This provides an ince-breaker to casually “friend” new casual friends. [FX: Cue all the stereotypical Myers-Briggs introverts grumbling into their coffee.]

    It would be a shame to lose that.

    This raises questions (especially in conjunction with Alastair’s mention of a FaceBook export feature,) about whether this meta-data information could be synchronised.

    (Another example, going further: Should it suggest to you a FaceBook friendship between you and everyone else in the same photo as you?)

  7. Don’t confuse this with Flickr advocacy, but I thought it might be helpful to quickly put together a matrix together with my best guesses about Flickr suitability.

  8. Wow! Thanks, Chris! You are at least two weeks ahead of me!

  9. Its funny because I was going through the same sort of ideas as you. I personally prefer Flickr due to many reasons, particularly licensing.

    However when it comes down to it, I KNOW for certain that my friends will not tag, comment or otherwise do anything social with those photos unless they are on facebook. This is I have a huge number of friends on Facebook, and almost no one on flickr. So there are two things

    – I need an OpenSocial deal where people with a FB account can see the link to the photos, open the photo site and start seamlessly commenting/tagging
    – I need flickr/(name your other ideal photo site) to send comments/tags and other meta-data BACK to my social networking system
    – From a marketing point of view, I need it to be well known that this photo site DOES operate like this with FB. Because I’d not risk people not commenting on the photos.

  10. From a marketing point of view, I need it to be well known that this photo site DOES operate like this with FB.

    Oh, I want better than that. Comments and tags made on FaceBook should share out to the other sites and vice-versa. I don’t need to do a bit marketing push if it is so transparent that it happens naturally.

    (It’ll never fly because it clashes with the FaceBook API’s privacy rules.)

  11. You are at least two weeks ahead of me!

    Sadly, not on comment grammar or editing.

  12. James: I have a similar discrepancy w.r.t. friends on Flickr and Facebook. However Flickr has a Guest Pass feature which will let you send a link to people which will allow them to view the photos without (necessarily) joining Flickr.

  13. Re definitions of portrait: I was aiming for some distinction between candid they-don’t-even-know-they’re-on-camera photos and formal set-pieces as I thought I’d seen you favour the former over the latter. If that’s no longer the case, then no matter. There’s another extreme to this, with a set, props and layers of grease for the camera soft focus. I’m pretty sure you’re not collecting these requirements for this last kind of photographer.

    Those crazy kids with their camera phones might take lots of photos, but they keep far less – and certainly don’t publish as many. How many photos would you expect to try and publish from a single event? Is it less than 10? 100? 500? 1000? I’m thinking only those who know they’re getting paid for their photos would be over the 500 mark, simply because of the amount of time it takes to move between worthwhile shots and any sane person would get tired of taking so many shots.

    Your feature of not overwhelming people browsing collections with this volume of photos would be key here, especially if you’re linking multiple photographers work for the same event together.

    Also, last I checked camera phones can’t match the quality of a half-decent SLR, thanks to the interference that’s introduced by camera phones using ever smaller CCDs (one place where more pixels != better photos), plus their inability to add good lenses with optical zoom or allow control over depth-of-field. It may be some time before any of these are addressed in a camera that is also a phone, let alone a phone with some extra space on the back.

    Another side effect of not trying to be pro about it is you don’t need to store the super high-res originals, but can stop at breathtakingly large photos for printing on A4 or smaller. Or, maybe you just mean to make available (to who?) old skool print photo size.

    Do you want to store the RAW image file, or only JPEGs? How large can a RAW get?

    Lastly, do you want to be able to give away ownership of some photos too, or otherwise mark it for Creative Commons usage? (I note this because it is subtly different to giving ownership to another person – there’s the possibility of attribution clauses, etc.)

  14. How large can a RAW get?

    Pretty large I’d say. My 40D produces RAW images that are about 13M, versus 3.3M for the JPEG version.

  15. >Since around November 2007, when an event is organised on Facebook, I have put up the photos of it there

    But a Google
    photos site:facebook.com
    finds four unrelated sites.

    On facebook.com a search
    photos
    gets
    Found no results for …

    So I can’t find any alibi-proving of myself at your last (?) shindig … and that make me sad

  16. This facebook don’t work proper …
    In the above, for
    photos site:facebook.com
    read
    angle OddThinking’s surname angle photos site:facebook.com

    and for
    photos
    read
    angle OddThinking’s surname angle photos

    And don’t get me started on the usability of having the Preview button below the Submit Comment button

  17. Geoff,

    I see three complaints here.

    The first complaint is that Google can’t find my photos on FaceBook.

    I am happy about that! That is a privacy feature.

    The second complaint is that Facebook can’t find my photos on FaceBook.

    That search box isn’t Google-powered. 🙁

    Your choices seem to be:

    1) First, find the photos application (either with the search box or under Applications.) and the page through your friends albums looking for interesting ones.

    2) First, find the person, and view the photos OF THEM, and hope that leads you to interesting albums.

    3) First, hope that they tag you in the picture, and then visit the photos of YOU!

    (In any case, I didn’t take a camera to my last shin-dig, and the photos that other people took, I didn’t see you in them. No alibi for you!)

    See my FaceBook evaluation for more comments about the photo features of FaceBook.

    The third complaint is that Preview should appear before Submit Comment.

    Interesting point. It isn’t an option that I can configure, but maybe worth escalating to the plugin author. Thanks.

  18. >Google can’t find my photos on FaceBook.
    >I am happy about that! That is a privacy feature.
    OK

    >The second complaint is that Facebook can’t find my photos on >FaceBook.
    >Your choices seem to be:
    >1) First, find the photos application (either with the search box or >under Applications.) and the page through your friends albums >looking for interesting ones.
    Easy when you know how. Or have bothered to look.

    >(In any case, I didn’t take a camera to my last shin-dig, and the >photos that other people took, I didn’t see you in them. No alibi >for you!)
    Weird thing is … my home photo is from your shin-dig. I didn’t take it. And I didn’t set it as my home photo. I do recall a monstrous Canon taking photos … but who could set my home photo? I see that you tagged me in a photo. Hypothesis: this was the first tagged photo so FB makes it my home.

  19. I was aiming for some distinction between candid […] and formal […] I’m pretty sure you’re not collecting these requirements for this last kind of photographer.

    Interesting… It is true, that I am more concerned with the former, but I am at a loss to how it would affect the requirements.

    Those crazy kids with their camera phones might take lots of photos, but they keep far less – and certainly don’t publish as many.

    Ahh, but maybe that’s because they don’t have the right photo publishing software! Then again, it could be because the photos those phones take are rarely worth keeping.

    How many photos would you expect to try and publish from a single event? Is it less than 10? 100? 500? 1000?

    I have taken well over 1,000 photos for some (multi-day) events. I would publish 100-200. More typically, I would take 100 photos and publish 25.

    I’m thinking only those who know they’re getting paid for their photos would be over the 500 mark, simply because of the amount of time it takes to move between worthwhile shots and any sane person would get tired of taking so many shots.

    Clearly demonstrating that I am not sane.

    Your feature of not overwhelming people browsing collections with this volume of photos would be key here, especially if you’re linking multiple photographers work for the same event together.

    Yes, agreed.

    Also, last I checked camera phones can’t match the quality of a half-decent SLR […] It may be some time before any of these are addressed in a camera that is also a phone

    Yes, the camera phones are much poorer quality than an SLR, and I doubt that miniature cameras will ever reach the quality of the much larger equivalents, purely because of lens quality (unless there is an innovation that is far more revolutionary than merely waiting for Moore’s Law.)

    Another side effect of not trying to be pro about it is you don’t need to store the super high-res originals, but can stop at breathtakingly large photos for printing on A4 or smaller.

    The file-size (at least, for JPG,) is going to be largely determined by the mega-pixel count (and hence, larger for newer phones, than the older DSLRs). As for the originals; I have storing archival quality as (merely) desirable, and storing to A4 quality as important.

    Do you want to store the RAW image file, or only JPEGs?

    I didn’t think to specify file formats. JPEG was all I needed. Other formats are just bonuses.

    Lastly, do you want to be able to give away ownership of some photos too, or otherwise mark it for Creative Commons usage?

    Good point. I worded it as “Attribution of copyright details” in the cleaned up version which I hoped would cover that. Being able to contact the owner is the bare minimum. Being able to tag particular images as open for copying would be nice too. My personal copyright clause is fairly open and yet subtle.

  20. Re formal photos: if you cared enough to worry about what jaunty pose your subjects stand in for the photo, you may also care what arrangement these photos have on the page – you wouldn’t want to ruin the aesthetic you’ve been going for. Deliberately not targeting this kind of user avoids interactive layout issues like this, and may avoid all layout issues (although I bet you’ve got more to say around this too).

    I have taken well over 1,000 photos for some (multi-day) events.

    Damn, somehow I’d completely forgotten about multi-day event photo bonanzas. Here I was thinking the longest anyone would hold a camera for without remuneration was 6 hours. I’m very sorry. No really 😉

    Oh, more questions: I’ve noticed that you want it to read your date and time info from the EXIF, but what if its wrong? Is that one of the little things that’ll need retouching when it’s added to the software and then resynced with your private copy?

    Is timestamp and location info (e.g. geocode) something you’d want to keep private, or selectively publish?

    Will your next camera have a GPS built-in?

  21. Oh, more questions: I’ve noticed that you want it to read your date and time info from the EXIF, but what if its wrong? Is that one of the little things that’ll need retouching when it’s added to the software and then resynced with your private copy?

    Aperture handles this case elegantly: you basically tell it the camera’s timezone and the photo timezone on import, and it works out everything else. (No idea if it can handle the camera-lost-time-reverted-to-factory-time case also, but wouldn’t be surprised)

    Frankly my initial exposure to Aperture has consisted mainly of little revelations like this; requirements I didn’t know I even had. They have obviously looked at professional photographers and concentrated on providing a smooth workflow for importing, categorising, editing, publishing and archiving images. And dabblers like me reap the benefit. I shall blog more about this, but even after a short time I am sold on the benefits of this tool and those like it.

Leave a comment

You must be logged in to post a comment.

Web Mentions

  1. OddThinking » Online Photo DB: Stage 4n - Evaluation of Flickr Pro

  2. OddThinking » Online Photo DB: WordPress mini-evaluation

  3. OddThinking » Online Photo Database Project - Status