February 06, 2014

Think On My Sins: Configurator & the simplification of Adobe tools

I fought the sprawl & the sprawl won.

I always intended to do a long series on what I’ve learned from failures, yet this will be the second & final installment (a bit of a meta-failure). Well, take it for what it’s worth.

In my many years working on Photoshop, I was sort of obsessed with the app’s inexorable growth & complexity. For example, in “Psst–wanna see Photoshop 15?” (Oct. ’05) I talked about the rate at which menu items were getting added. Even if the team somehow found a way to *drop* 60-70 features per release (impossible), we’d only tread water in terms of complexity.

To make real progress, I proposed breaking Photoshop into task-based chunks (for example, showing only photography features when you’re working on photography). Thus you could really feel like the app was made just for you, and that it revealed exactly the right set of features (and tips) just when you needed them.

I didn’t trust Adobe, myself, or any top-down approach to get these chunks exactly right. Instead I proposed letting customers tune the app themselves, building their own workspaces which combined layouts, menu setups, and keyboard shortcuts. Critically, these workspaces could also include custom panels—layouts that you could Lego together to fit your exact needs.

Enter Adobe Configurator. It offered a simple set of building blocks, letting you mix together a custom panel from any combo of Photoshop tools & menu commands you’d like. I never expected most users to invest the time—maybe 1 in 100 would, I figured—but I hoped that a small number of thoughtful, motivated users (the sort I once was) would create & distribute stuff for everyone else.

So, what happened?

  • Configurator gained a couple hundred thousand downloads—pretty great for a nerdy utility posted to Adobe Labs.
  • Some authors like Vincent Versace created & distributed custom panels.
  • The Photoshop team make Configurator much more powerful & used it to create the Knowledge panel for CS5. When you’d click a particular workspace (e.g. 3D), you’d then get a grouping of relevant tools plus interactive How-To content. It was pretty damn cool, if I may say so.
  • Most people didn’t do much of anything, however.

What went wrong? What can we learn?

  • Sharing custom panels was far too hard. (I won’t describe all the onerous steps for packaging, decompressing, etc.)
  • The Knowledge panel didn’t ship in the box with CS5 (the whole 64-bit/Cocoa transition was dicey enough that we had to cut it at the last minute), and the team never included it later. People didn’t care about in-app help, at least to anything approaching the degreed I’d hoped.
  • In making Configurator support this sophisticated use case (i.e. “eating our own dog food”), it became complex & intimidating, when it should have erred on the side of simple on-boarding.
  • Ultimately, the whole problem reminds me of dogs chasing cars: What would they do if they caught one? That is, everyone likes to bitch that apps are too complicated, but when you give them the chance to streamline & reorganize the UI to their tastes, they don’t know what they’d do differently, or they just don’t care to bother.

Could things change? Perhaps:

  • The “settings sync” feature introduced in CC could morph into settings sharing, letting me make what’s mine yours & vice versa. (Example: I go see Michael Ninness teach “Photoshop for Web Design.” I type “ninness” into my copy of Photoshop, see Michael’s custom workspace (including Configurator-style custom panels that present tools with context), hit “ok,” and have it all on my system, period.
  • Adobe could create a Tumblr-simple publishing system for people to share their interactive how-to content, making it appear right within CC apps. (I naively thought that authors would see Configurator’s ability to include HTML views & immediately start populating them. I came to realize that traditional authors are used to writing a manuscript, sending it off, and receiving cash—no futzing with the mechanics of printing & distribution.)

Ultimately my whole obsession may have been a fool’s errand. You don’t turn an apple into an orange; you just make new oranges. Photoshop, Illustrator, InDesign, etc.—maybe they just are what they are (the ultimate in power and control rather than approachability), and nothing will or should change that. Instead Adobe should build fresh new tools that complement, rather than seek to replace, these powerhouses.

At this point the future belongs to you & to the teams at Adobe. If this stuff is important to you, please let them know what you need & want, and why.

Thanks for reading,

Posted by John Nack at 10:25 PM on February 06, 2014


  • Guy Einy — 10:40 PM on February 06, 2014

    Thanks for sharing John, what you describe exemplifies one of the most frustrating aspects of building things – nobody really uses it. Not the way you intended them to, anyway.
    You try and solve a real pain, you succeed, and yet fail.

    P.S. Photoshop 15 is just around the corner… :-)

    [I know!! :-) I wondered if others would notice that. –J.]

  • Max Dunn — 11:01 PM on February 06, 2014

    Configurator was an awesome utility, but it wasn’t deep enough to do anything on the level of CS Extensions, yet theoretically it could have enabled mere designers to build truly powerful extensibility. I doubt they went there in droves.

    CS Extensions had a similar flaw: they actually *did* provide a power that is possible only with true programming chops, so a much smaller number of users but a much greater potential end result in terms of what was created. CS Extensions probably got greater use, because it is us programmers who fixate on extending InDesign, designers tend to be more like “how do I maximize the base product” for whatever reason.

    Both are victims not only of the general tendency of designers and programmers NOT to extend CS/CC applications, but also of the death of Flash.

    Because they were both Flash-dependent and Adobe took the odd decision to wipe Flash right out of CC instantanously without a good replacement anywhere remotely on the horizon (two replacements: the internal regression to C++ and the marketing message that a non-existent HTML5-ness will appear, coupled with a beta proving it’s decades away).

    No wonder you went to Google, that’s where the new cool part of Photoshop comes from. Extensibility is only for the rare few, but it *would* and *should* have been for everyone.

    I love the vision of Configurator and I hope you can attain something at Google that does put such power in the hands of mere mortals. I am afraid, however, that both Adobe and Google offer constraints at a bureaucratic level that you will end up with more failures to document. Hope I’m wrong, but very thankful that your “failures” express vision that would truly be the best path forward. A Photoshop/Illustrator/InDesign that can easily and powerfully be extended by normal people remains a compelling vision.

    • John Stevenson — 7:23 AM on February 07, 2014

      I agree with this, pretty much from top to bottom. However, there was one flaw in the overall scheme which isn’t covered – it wasn’t good to try and bring simplification to the user experience with a framework itself which was way too complex. CS Extension building was made possible for multiple CS applications, Photoshop being just one of those. So the whole extensibility platform was compromised from day one and (subsequently) the support effort too fractured.

      I don’t write code, and (therefore) loved the concept of Configurator. But when I found that Panels constructed by and within Adobe as demo exercises didn’t always run as expected and used scripting, well, that was a big disincentive. There were even Tools in Photoshop (fundamental ones, like Brushes) which couldn’t be scripted anyways … So, overall, I concluded that a helper app – one which allowed for the publication of tutorial content along with customized toolkits – would get good traction, this particular platform was hobbled from day one.

      Finally, I don’t see how the “Adobe could create a Tumblr-simple publishing system for people to share their interactive how-to content” would be any better. The Marketplace still exists: http://www.adobe.com/cfusion/marketplace/. Folks still contribute resources there (though the banner hasn’t been updated since CS6 … http://tinyurl.com/d7m6hz). Adobe could do (much) better, for Photoshop at least, by revisiting that, making it link to/from Behance, and paying more attention to user inputs from here: http://feedback.photoshop.com/photoshop_family. Interestingly, Revel (formerly Carousel) has now become the dominant topic in questions and topics posted there (which is a lesson in itself on current needs and future trends).

      • John Stevenson — 8:07 AM on February 07, 2014

        Just out of curiosity, I looked at the newest addition to Adobe Marketplace. It is a drawing effect Action for Photoshop (heck! – that’s new …?) – posted a month ago. “No skill is required – just 1 CLICK”. $20, and 1,734 downloads to date.

  • Omke Oudeman — 4:45 AM on February 07, 2014

    I was one of the enthusiastic downloaders/testers of Configurator and still find it very sad it never get developed to what it promised to be. My highest need was a customizable action panel to reorganize my ton of actions using button mode.

    The promised drag and drop from existing action never get realized and the option to create the buttons yourself to a reasonable small size was (albeit a bit painful) nearly impossible due to the non removable action icon in each new field.

    Luckily the button mode is still present in current action panel (for how long??) but we still have no control over it, the resizing of the panel decides how many columns we get without us being able to pinpoint it as we like.

    So I’m left with a lot of scrolling through the current action panel. Yes, it was very sad Configurator never went to that what it could be.

    But to be honest, it is not the only feature that has gone this route. Think about Application Frame (could be marvelous, still is a bit clumsy) think about Bridge (for me still unmissable for my workflow even in it’s current state. If only 25 % of all filed Feature Requests was fulfilled Bridge could be fantastic but since CS4 it is treated as Adobe’s non wanted step child), Mini Bridge (promising in CS4, totally screwed in CS5 and still by far what it could and should be…). Will they be trashed in the near future also??

    I really hope you are right in stating the future belongs to you (users) and the Adobe teams, but at this moment I have the sad feeling that it is very less me, a bit of Adobe teams and very much management decisions (read marketing decisions) that will plan the future.

    I would love reading from you to proof me wrong :-)

    Meanwhile thanks for all your contributions over the years and all the best in your new job!!


  • jlua — 4:59 AM on February 07, 2014

    In my humble opinion, one way to simplify Photoshop at each new release could have been to replace old relics as new improved functions get introduced. I think I understand that Adobe tries to be very conservative in eliminating “relics”, because I suspect there will always be groups of users who will complain. But there are plenty of prehistoric “relics” lying around everywhere in Photoshop, that add complexity and could be very well be eliminated as their functionality gets improved and replaced with new functions.

  • Keith — 5:43 AM on February 07, 2014

    I just kinda forgot about the Config… but I did tinker with it. And was thinking just the other day wouldn’t it be nice to have just a line up of the tools I use (- the one’s I never use nor know what to do with really).

    And as a Ps tester-outter (thanks John) it was amazing to hear the howls around the world from folks who’d get so disappointed if a tool was changed, removed or even its icon changed.

  • Chuck — 6:00 AM on February 07, 2014

    You got me testing Configurator back with CS5. I thought it was very cool, but it lacked what I needed it to do: be more interactive with extendscript. What you mentioned about publication of panels is true. Most of us use scripts and panels for our own use, so basing use on the numbers of panels published will give you a very low inaccurate measure of Configurator’s use. It was a great idea, and still is. Hate to see you leave Adobe, but I’m glad you’re looking forward to the new opportunities at Google

  • Edward Caruso — 6:45 AM on February 07, 2014

    I dabbled in Configurator but found that designing my own panel alittle was a bit frustrating and more work than I expected. I saw the utility and potential but I felt that the instructions were alittle lacking. Mainly I could see it took time to get it right (and looking right) and I had other tasks pulling me away. Kind of how I wish I could learn PS scripting one day. There was the suggestions that users could share panels (especially PS authors/experts out there) that we could download but that really never took off.

    John best of luck at Google (maybe you can suggest improvements to the camera in my Nexus?) and look forward to following your next blog.

  • Jeremy — 7:00 AM on February 07, 2014

    I always thought Configurator was a good idea but I disagree with you on one point: not trusting Adobe to build the workspaces themselves. I think users generally don’t like to spend much time tinkering with their UI and want to just create, and if Ps shipped with several UI configurations via Configurator then it would have been more successful. I think Adobe’s default workspaces for many of its products have been well-done. Of course, Configurator would have allowed power users to roll their own and/or modify Adobe’s defaults, but many creative users wouldn’t have gone to the trouble.

  • Stephen Walker — 7:53 AM on February 07, 2014

    I love Configurator and have custom panels for PS that I’m really going to miss one day. They make life so much smoother.

    Just a note – not everyone who uses Configurator posts panels to share. I’m sure no one will be interested in mine as they are quite specific to my workflow.
    So maybe the apparent lack of use is not the true picture.
    My only complaint is that to change something you need to go back to Configurator and alter and re-export.

    Interesting to see Illustrator has the ability to drag and drop tools into a custom Tools Panel. Works very easily – pity it’s limited to tools.
    Put its ease of use together with Configurator and make it a CC download and I think it would have been a hit.
    Oh well. Not the first thing I’ve loved that’s vanished ! ;o)

  • J. Peterson — 8:05 AM on February 07, 2014

    Excellent analysis. If a feature ever flops again, consulting fees await : ).

    [Oh, let me know when you’re ready to revisit non-destructive filtering. :-) –J.]

    At this point Photoshop’s UI seems like the piano keyboard; once a sufficiently large population has climbed the learning curve, there’s little sense in re-arranging it.

  • Arnon — 8:10 AM on February 07, 2014

    Configurator was great. It’s a shame you dropped it. 200K users is not bad at all. Even if only 20K might have actually used it. It was 20K power users. Those are your influencers. Those are your promoters.

    Configurator does not look lke such a high maintenance feature.

    [Surprisingly it is, in that Flash panels are going away. And it *really* used to be high maintenance, as I had to hand-record/format/paste every one of 1000+ commands into a big spreadsheet. Thank God that finally got automated. –J.]

    I don’t really know, certainly. But for me, it allowed me to have simple and useful UI customizations that worked well for me and saved me many clicks.

    Just because the people who used it didn’t share the features (because it was hard to share, and maybe the customizations were very “personal taste” things) does not mean that people didn’t use it to good effect.

    [Certainly some did & continue to do so, but usage stats tell us that fairly few did overall. –J.]

  • Allen Cobb — 9:36 AM on February 07, 2014

    I love Configurator, but I have trouble forcing it to do what I need, it still isn’t supported by Ai, the mindless housekeeping isn’t sufficiently automated, and it’s been hard to publish or monetize any of the grand plans I had. So I use it for consolidating some tools and supporting some very specific workflows, but it languishes — quite unnecessarily.

    There’s a significant disconnect in the way most companies interpret “results” in the marketplace. Look at the hundreds of MP3 players and PDAs in the market before the iPod & iPhone. It was very clear that neither type of product was a killer, and the idea of jamming a phone into a PDA or a player was dismissed. Why would everyone want all that in one device anyway? A dedicated phone is better. I don’t need thousands of tracks on my phone, either. A PDA is a nice idea, but only techies and certain kinds of salespeople would *need* one.

    And yet all of that conventional (fake) knowledge went up in smoke when Apple reached critical mass in features, functionality, usability, and design. Suddenly, an MP3 player PDA with a phone jammed in is the bees knees. Why? Because it’s not a phone, or a player or a PDA: it’s a personal computer.

    It should have been obvious, but how many strategists could guess that all those sensors and features would make something that was irresistable? It would have been just as likely that the *same technical capability* just turned out wildly geeky and unmarketable.

    So I hope Adobe doesn’t read too much into the present state of acceptance of Configurator. It’s hugely misleading. There are tons of things that a “more configurable” Photoshop, InDesign, or Illustrator (please!) could do to facilitate the apps or even to directly increase sales. Oranges aren’t needed, just that elusive critical mass of functionality, ease of learning & use, ease of distribution & optional sales, and a dash of elegance. To get there, one additional ingredient is an absolute requirement — stepwise refinement.

    Meanwhile, maybe Google will come up with a Universal Configurator. It could be done… OTOH, even Google has a weird track record with, uh, stepwise sudden disappearance.

    But this is all for your successor (an impossibility). So again, bon voyage!


  • Mark Maio — 9:57 AM on February 07, 2014

    Let me start out by saying I approach this from two halves of my photographic life. I am a fine art photographer who after thirty years of using film embraced digital imaging and use PS and LR as my digital darkroom. In no way am I a power user nor do I have any programming skills. The other half of my life is imaging in science, medicine and research. When PS CS3 Extended was being developed, I was one of twelve individuals chosen by Adobe to form their Biomedical Imaging Advisory Group. As a result of this I started the Digital Imaging Institute for Science and Medicine as a vehicle to teach basic (and scientifically correct) uses of PS and digital imaging to these fields.

    I was excited when CS3 Extended and the initial version of Configurator came out. Most PS users in the medical, science and research fields were using old versions of the program (PS6/CS) and had no idea of the advancements in the program let alone the features of Extended. They also had no time to learn how to become a PS expert. They just wanted to learn the features of the program they needed to do their imaging work. For the majority of these people, when they opened the new versions of PS it was the “deer in the headlights” look. They had no idea where to begin or where, even after going through a learning session with me, how to find the tools they wanted to use.

    When Configurator was announced, I thought my problem was solved. The problem I had was that it wasn’t intuitive for me to learn how to create panels. The instructions I could find on the Adobe site were okay but not being a power user/programer, I had questions that I could never get answered. Something I expected, given its intended use, should be easy to use, wasn’t. It just seemed like the program was put out there and sort of left alone. In a way, it was like what happened with the medical/scientific tools in Extended. They haven’t really changed since they came out in CS3 and there hasn’t been any new features added to that section.

    John, it seems like each product team is so busy that you get out new features and then have to start working on the next new features before the end user starts to use those you just developed. In the end parts of the program just sit there without anyone given the time to move it forward. If the new feature isn’t one of those “wow” tools that the general photographic community embraces and starts to teach, lecture on and write books about, it dies a slow death on the vine.

    In a sense my feeling is that products are developed with the attitude of “If we build it they will come”. I saw this happen with PS Extended. It seemed like the expectation was that because there was this new version of the program that addressed imaging in medicine, science and research, this market would some how just “know” about it and buy it without any real marketing efforts. I think the same was true of Configurator. In the years after it was released, at each of my lectures or classes, I asked if the attendees knew about it and to a person no one had.

    So no, it wasn’t a sin to develop it. It was a sin that there weren’t the resources put into making it easier to understand how to use for the average person and to advertise it to the people who really needed to know it existed so they could find out it was actually available.

    Best wishes with your new position.

Copyright © 2021 Adobe Systems Incorporated. All rights reserved.
Terms of Use | Privacy Policy and Cookies (Updated)