512x512 Icon

Quite the opposite. A user can perfectly say, “DT is ugly”. A critic doesn’t need to perform better in what he/she criticises about the object of his/her criticism. Especially not in such a consumer/vendor relationship. Unconstructive criticism is perfectly viable. And devonthink gets the message: “Upps, users think our software is ugly.” Or now, in post-1.5 and beta2-times: “Hm, users still don’t consider DT as pretty. At least, users think that it’s not as ugly as before. They think it’s a bit inconsistent, dull, unpolished.”

The devontechs can then ask themselves whether they would want to invest time and ressources in a major UI redesign. Whether they would want to study the fanciest software out there, whether complex software has ever been attributed as “pretty” before, read journal article on UI design, surf UI-designers blogs, study complex software GUIs and what else. They could certainly just leave everything as it is, as DTs functionality is second to none and the UI is OKish, so that most users around wouldn’t stop using it, anyway.

I’ve talked about this before on the Forum and so will not provide examples of decent interface design. The fact of the matter is this: one shouldn’t actually notice good interface design. It is good because it doesn’t get in the way. I’ll second MDAnderson. DTP is a best in class information manager. I’ve used them all : Yojimbo, Together, Evernote, you name it. DPT is simply better, more sophisticated and capable of more (although I’d rather DPT 2 had come through as an indexing service of all user files/ folders rather than as a self-enclosed db - but hey, that’s for version 3).

HOWEVER, I dread firing DPT up as its poor interface actively subtracts from its usefulness. I’ve been disturbed to hear some DT staff give the impression that they know better than Apple when it comes to interface design. On the evidence of UI design undertaken by both companies, I’d beg to differ with the DevonThink team.

That’s a bit like saying "It’s easy to make a beautiful male or female model more beautiful. It’s much harder to make an ugly male or female genius beautiful without affecting his or her I.Q. There is some truth in the statement, but they both use the same makeup; one just applies more of it and in slightly different ways.

Not a single piece. But I would also not sacrifice a morning in Paris on a gray, drab, and cold day for the sunshine and superficial warmth of Palm Springs (apologies to any friends I have left in the Coachella Valley). I’m fairly sure, however, that the French Tourism Board isn’t going to send me brochures full of drab, uninspiring pictures to try and sell me on Paris or convince me to tarry there for an extended period. :confused:

The Paris Tourism Board has no control over the weather, and not much control over the things that make Paris worth visiting. All they can control is when and where they send the photographer. If they want pictures with sunshine, all they have to do is wait.

DevonTechnologies, in contrast, has complete control over both the UI and the functionality. That is a blessing but also a curse. They can improve the UI whenever they want. They can add functions whenever they want. They can ship the product whenever they want. BUT every hour spent improving the UI is an hour during which the product is not shipping. Every dollar (or euro) spent on UI designers is a dollar not spent on algorithm developers.

If you want sunshine, don’t visit Paris in February. But if you want to see the Louvre, and February is the only time you can go, it seems unreasonable (though very human) to complain that the weather is nicer in Palm Springs.

Katherine

Then to extrapolate from your argument, if one wants an attractive UI don’t buy DTPO because there will never be the right season? Given that I prefer to see if I can convince the weather gods at Devon-Technologies to part the gray clouds a bit. As you say, they have more control than the Tourism Bureau.

Exactly, though, with your analogy in mind, my complaint would not be that the weather is nicer in Palm Springs, but that the curators at the Louvre didn’t make an effort to mop old mud from the steps, switch on brighter entry way lights, and dust off the paintings because, of course, the Louvre has the best art collections in the city…and that should be enough. Besides, we have 20 crates of artwork in the basement that has to be unpacked, restored, and displayed.

which is why no one 512x512 icons :wink:

Indeed. :smiley:

I love this forum. Thank you all, especially Katherine, for an enjoyably diverting discussion.

Bonne chance avec votre écriture.

I absolutely adore the icon. I think it’s beautiful. I don’t know how many pixels it is or what it is supposed to be, but I love the design. I think it’s the reason I chose the product.

The application itself is a tad awkward, but I am sick to death of beautiful looking Mac Apps that do next to nothing. All form, no function. Too many Mac developers are hyper-focused on design and ease. Everything else of importance I’ve had to go back to the hideously ugly world of Windows, because I’d rather have something ugly that does what I need than something pretty that doesn’t.

First, I use DTP daily. I love it because of its power. I hate because I just hate its user interface design and poor usability.

My guess :

  1. DTP staff are very talented programmers, and very cultivated people.
  2. They don’t have designers into the core of its team.
  3. They simple don’t care about design, they think design is accessory.

I could not agree with the idea that design is accessory, but I can live with. It is difficult to find a team like Panic, where functionality and design have the same priority (I think that of the most powerful and useful apps in the Mac platform is Coda, still it has one of the better user experience and user interface design). But, I can’t forgive usability mistakes as those present in DEVONthink.

Using DTP is just a rational act where you got to know the size and disposition of every window, panel and navigation state, if you want to be productive. I simply renounced to every view different than “three panes”, and I use DTP as if it were Mail.app. Changing views means losing navigation state, losing document focus, and losing column sizes. And, as I always said, there are SIX views, plus “widescreen”, coverflow, drawer, info and groups panels, etc. I think that it is too much, and complexity needs very good design.

Thanks.

While one or two - and only one or two - of Dieter Rams’ principles of great design don’t apply directly to software, I think they’re worth recalling in the context of UI design. Rams is the spiritual father of a number of Apple’s hardware and software (e.g., iPhone calculator) designs.

vitsoe.com/en/gb/about/dieterrams/gooddesign

I think radii0 hits it on the nose. DTP needs an UI worthy of its functionality. It doesn’t have that and I’d like some sort of position statement from the DevonThink team about the future of the program’s UI. I’ve paid my money for 2.0 and I’ll pay for 3.0 too - but only if the UI improves in the short term…

I don’t know what everyone else is using but I got the new iMac 27" for school, it costs less then $1800 and has a 27" screen which has the same resolution as the old 30" cinema display. I think snow leopard uses 128 icons as a default but all my icons are at 256 by default now.

I like the colorful icons and there is still plenty of room left over.

Not sure how it contributes to your design process but bigger icons are not some crazy thing that nobody uses it’s becoming more common I think, the prices for big monitors have dropped a lot. The iMac also lets me plug in a macbook and use the iMac as another big screen.

Anyway the 128px icons do look blurrier then all the other ones and it does noticeably show up.

That’s all.

@ radiio & drspk:

Let me start with a disclaimer: I’m not one of the developers. I came to DEVONtechnologies as a very heavy user of DEVONthink databases, and that’s still my perspective.

True, there are 6 views into a DEVONthink database, and some of them are modifiable. That contrasts strongly with Panic’s Coda application for one-window web development.

I spend most of my working time in DT Pro Office 2 inside the Three Panes view, with the Sidebar showing, but not with the Widescreen or Cover Flow options active.

Once in a while I switch to the Split View, for one of two reasons: I’ve opened a group in its own window and am using that for the larger size of the document view on my 13.4-inch laptop, or I wish to quickly select ALL the content of the view, even if I’ve violated the organizational guidance that there should not be unclassified files at the top level.

I rarely use the List or Icon views, although the Icon view comes in handy if I’m looking at a group with graphic elements, especially photos.

But I will now introduce a strong case for a view alternative that does NOT result in display of a selected item. DEVONthink uses Apple’s PDFKit to display PDFs, and WebKit to display WebArchive documents. If I’ve captured a strange ‘flavor’ of PDF that crashes PDFKit, or a malformed WebArchive that crashes WebKit, that document cannot be deleted when selected in a view that attempts to display it; a crash will result. Solution: switch to List or Icon view, select the item and delete it. That works, so an alternative to views with automatic display of selected items is useful.

There’s another reason for some of those views, which is simply accommodation of user flexibility and preferences. There are power users who prefer the stark simplicity of the List view, or the Columns view for working in complex hierarchical structures. Although in the current public beta 7 the List, Icon and Columns views do not automatically display a selected document, if the Space bar is pressed a Quick Look view will open.

There will be some changes in the forthcoming public beta 8. The Columns view has been modified to provide a display of a selected document (it’s now the favorite view of some users). And tagging is introduced, so the Tags view is activated. By default, all existing groups in the database will be listed as tags. However, any or all existing groups can be excluded as tags. Click on a tag in the right column, and a list of its members is presented.

Another significant feature in pb8 is that one can open documents as tabs in a view window, move back and forth between Split and Three Panes views and even move among databases, while retaining the set of tabbed documents — even adding tabbed documents from different databases and from See Also suggestions. Although such a tab set cannot be permanently saved (at least currently), one can create a rich text document holding links to them, and so recreate the set at any time.

OK, I’m begging the question. Can the user interface of the DEVONthink applications be improved? Certainly. Does that necessarily imply that a single window interface would be an improvement? I don’t think so. Users do much more complex things inside DEVONthink than in Coda, with a variety of personal preferences and workflows.

There’s no question but that the design of the DEVONthink applications have been engineering and functionality driven. Different users use DEVONthink differently, and the developers have taken input from a broad spectrum of users. The applications have therefore evolved so as to handle (at least to some degree) the document management/information analysis needs of diverse users. There have been many comments to the effect that the DEVONthink applications are at the leading edge of power, compared to other such applications for the Mac and Windows.

DEVONthink has been around since 2002. Today’s editions are very different compared to the original. Throughout the history of the application there have been complaints or suggestions about user interface issues. Some have been satisfied, others, not. Things that some people consider simple and elegant are viewed as ‘butt ugly’ or unusable by others.

I liked some (but not all) of Dieter Rams’ designs (yes, there are divergencies of opinion about design). But I’ve got to note that his design of the radio or the tape recorder wasn’t restricted to a single button or knob, unlike the ‘no interface’ design of the speaker. A person unfamiliar with calculators would need a good deal of instruction to make use of all the buttons, and some would find their needs or preferences for some types of calculations unsatisfied.

Eric has noted that the first priorities in development of DEVONthink 2 are stability and completion of the planned feature set, and that as these come into place (stability is already here, in my opinion), polishing the user interface would increasingly become a priority. Many items have received polishing throughout the public beta period. That will continue.

If, hypothetically, a Dieter Rams were brought in to redesign the user interface, what do you think he should tackle first? And if he did that, do you think some users would object? (I’ve got an iPhone. As good as it is, there are some things that drive me nuts.)

Bill, thanks for your answer.

I got to say that I use DEVONthink mostly as you use it. Three panes view, no coverflow or widescreen. I try to don’t over organize things and just use the search feature which is great. Working with large corpus sets and great complexity can only be done with search. I just create what I call “directories” (mainly in Zotero right now), and when I want to search for related items, I just hit a search (cmd shift F) in DEVONthink and voilà.

I hope some usability errors can be fixed for the final version. That doesn’t deny my hope of some day having also a great user experience as I have using Coda or other Mac software. I guess one can’t have it all.

To answer your question: What Coda introduced is not just a single window interface (the user can always open as much as windows as she wants), Coda introduced a very well thought modal interface. The most important aspect of Coda is not changing “views” of the same information, but switching between of “modes” of work. You can be editing a page in code view and then switch to preview and adjust it, or manipulate your server settings using bash, uploading your site to FTP, committing your results to a CVS repository or just reading programming reference books. And this is done with 5 modes. This appears to be very simple, but it is because this simplicity has answered a very complex problem. The genius of Coda is to have glued all of the relevant contexts of work of web development in a single environment. And it works great. I’m a web designer (indeed I split my time working as a web designer and trying to finish a PhD) and I really do appreciate the concept.

Views themselves are extremely flexible, you can have as much as you want, because panels can be arranged in a kind of “lego” way.

I have already been thinking how DEVONthink Pro could work with a modal paradigm (inspired by Coda) and I arrived to the conclusion that this is possible. I’ve even created Photoshop mockups to illustrate my thoughts, but I’ve hesitated to share them.

The idea is to have 3 main modes: Explorer, Notebook and Search. All modes have different views configuration.

“Explorer” allows the user to navigate and manipulate large sets of documents in the database, mainly with a three pane interface (sidebar, list of items and viewer). The sidebar has two states, a sources state and a groups state. Sources shows current and recent databases, general inbox, favorites, etc. Groups state shows the groups tree of the currently selected database. The groups state doesn’t show child documents, just folders. The second pane can show items contained in a selected group by icon, list or column. A third pane is the viewer of a currently selected document. Viewer could be closed. Coverflow is optional and just opens in top (forcing the interface to go to widescreen). Info panel can be set to float or to dock to the right of the window.

“Notebook” allows the user to read, compare, write and annotate documents using a two paned interface (sidebar + viewer). Here, the sidebar has child documents (this mode is very inspired by Scrivener). One third optional panel could be an annotation panel between the sidebar and the viewer (as the comments sidebar in Pages). Info panel is attached to the right of the window, to easily add comments and tags.

“Search” is a mode who allows the user to create advanced searches and filter the results by facets or metadata. The idea is similar to the search window in DEVONthink Pro. The sidebar works like a filter pattern, where the user can narrow his search by adding or subtracting criteria.

Hope this ideas could inspire someone a little.

I think the basic problem with DEVONthink is that it does a LOT. I wound up using and sticking with it, because at the end of the day it IS an engineering-driven application and Christian is a remarkably gifted engineer/programmer, who has managed to maintain and evolve an extremely complex application which winds up intersecting with the OS in myriad places, and does an incredible job of dancing through the minefield and dealing with all the broken crap and problematic frameworks which are under Apple’s control, and cannot be changed/revised, until Apple gets around to dealing with it.

My databases MUST be reliable and they need to work. I have never found anything within the same class as DEVONthink which is not a client to some very large SQL back-end running Oracle (or occasionally MySQL). I have databases with tens of thousands of groups, and hundreds of thousands of documents, and DEVONthink continues to work… rapidly, and reliably. Which is a miracle in and of itself.

Try loading all that crap up, into anything else that’s out there. You’re going to discover that once you pass a few thousand documents, the application slows to a crawl and starts falling apart at the seams. They don’t scale, period.

Some of the complexity can be compared to the neverending religious war between the dual, classic *nix editors: vi and emacs. I’m not going to wander off on endless tangents, but to sum it all up: vi is very fast, small and simple, whereas emacs is everything and the kitchen sink, LISP, and some cast-off UFO parts that happened to be laying around. Doing simple things requires insane levels of finger-dexterity, hitting 7 function keys at the same time, and works better when you have a mouse with at least 9 buttons, 3 scrollwheels, a gear-shift, and foot-pedals.

B’okay, I exaggerate slightly, but not much.

The problem with all of DEVONthink Pro’s different views and options and features is: they’re in there, because somebody – usually a collection of somebodies – finds that particular feature extremely useful/possibly critical, to their continued use of the product. By simplifying the interface, you are inevitably going to wind up stepping all over something, that I (and a few thousand other people), are presently using, and find very handy. What I like, and use, depends on the database and what I’m working with; and that reflects my personal needs, which may not be the same as yours.

None of this is an excuse for poor UI design, and DEVONthink is gradually and consistently getting better. It is, however, not exactly the ideal Apple app, and probably not going to be winning any Apple Design Awards anytime in the near future.

On the flipside of all that: how d’ya feel about iTunes? This is probably the ultimate instance of an Apple application which began as one thing, and has gradually grown to encompass everything and the kitchen sink, doesn’t really use any Cocoa frameworks at all; it fakes everything and lives off in an alternate universe coded in C++ so it can be cross-compiled across multiple operating systems, and act as a hub for all the shiny Apple gear, you plug into piece of s–t Windoze boxes.

It’s very large, it has to do a lot, and it’s probably the one Apple application people complain about the most.

The question is: how do you evolve all the power and complexity that’s present in DEVONthink, into a better user-experience, without taking some of the power away from people who want/need it.

Possible answers are: something along the lines of what Adobe has done with implementing “Workspaces” and further refining them in CS4, so you have the basic “Image Editing” workspace, the “Video” workspace, the “Pre-press” workspace, etc, etc, etc. While that’s a step in the right direction, and perhaps DEVONthink could offer a “Creative Writing” workspace, a “Scientific Monographs” workspace, a “Whatever Else” workspace, and allow people to create their own default workspaces, it’s still not going to simplify the program down to a level that’s really intuitive.

Preview is a really nice, handy, constantly-evolving Apple app. It’s great for reading PDFs, viewing and editing simple stuff within images, etc. But there is a lot it cannot do. Fire up Adobe Acrobat Pro 9, or Photoshop CS4, and … you have a gigantic, incredibly confusing – to the novice – array of tools presented to you. And Adobe has, for all intent and purposes, limitless money, to keep throwing a small army of UI designers at their product, in an attempt to simplify things.

Still, when the average person who has never used it, launches Photoshop, well… it’s not exactly Pixelmator or Acorn or whatever, but on the flipside, it does roughly 2000% more, if you know how to use it.

And we’re back to: figure out Final Cut Pro vs. iMovie. Different products, different audiences. Most people can intuitively grasp iMovie, without reading the help file, best of luck doing that with Final Cut.

So far as the Apple HIG goes: it’s just a lot of “blah blah blah,” the actual HIG amounts to: whatever Steve thinks looks cool that day. Which is why the Apple interface is pretty schizophrenic and has spent nearly a decade with a mis-matched collection of random themes – chrome, no chrome, dark brushed metal, blue scroll bars, no blue scroll bars, capsule-shaped buttons, no capsule shaped buttons, light background, dark background, the Pro apps which all use yet another interface altogether and are always in grey, etc – which is gradually moving towards some sort of cohesion and unity, but as of Snow Leopard, STILL isn’t there.

Having said all that, these are just thoughts and things to contemplate. I enjoy enriched environments, I like big colorful icons, I am sitting in front of an OS/X box, not running X + Enlightenment or Compiz Fusion, on Linux or BSD. DEVONthink does need to evolve the UI and address some issues, it’s just not as simple as it may appear to be.

One instance of an application that is comparable to DEVONthink Pro’s complexity would be Omnigraffle Pro. It’s nowhere near as simple as it used to be when it was the Lighthouse app Diagram! under NeXTSTEP (which Visio copied and ported to Windoze, and Microsoft eventually purchased), but it’s current incarnation has been constantly re-arranged, and the UI has evolved to still be relatively friendly and intuitive while offering a lot of power to people who need access to it.

Anywaze, just my thoughts.

Patrick

Thanks for this very articulate and interesting message on the topic. You’ve framed a lot of the issues really nicely and probably made me calm down a bit.

The only message I found truly annoying was the snotty attitude towards Apple’s HIG guidelines regarding a bigger icon. Big deal, it’s an icon, it’s not re-structuring the entire app and making it over, the snarky reply was really not called for.

Thanks again for your post and the interesting historical contexts, I had no idea about the origins of Omnigraffle or where it came from. I very well understand Devonthink is complex, it’s why I’m using it! I like the tools but do feel that maybe then workspaces for different purposes might be a good start.

How I feel about iTunes is that it’s a bloated horrid mess which badly needs to be renamed to something more appropriate and broken apart into more targeted apps somehow. It’s a music player first and foremost and it’s getting slower and more bloated with each revision.

Your point about Apple’s schizo UI design is also taken. Apple rarely follows its own HIG, I remember that back from when Safari first appeared. Chrome was for media apps, except for Safari because it looked cool and whatever else Apple felt like chroming, eventually the finder itself. They don’t seem to follow their own HIG very often but do a great job of the same function keys work the same almost everywhere, except when they don’t. Nevermind that comparison I guess :open_mouth:

There’s nothing snarky about my pointing out the HIG for icons.

They are Apple’s recommendations and guidelines. With things like Coverflow and a number of other Finder views, there is a marked difference in appearance between a 512x512 icon and a 128x128 icon. It is readily and easily apparent. So Apple recommends this to developers for a reason… there are a number of instances now where the OS will display an icon and if it’s not properly sized it will look like crap.

So that’s fine if you don’t mind the icon looking like crap, but it’s not like people won’t notice. But a substantial part of the Mac in general is attention to detail. They asked “Why should we?” and I think “Because Apple recommends you do so” is a fair response.

Nice read. Thanks, Patrick.

Hi I wasn’t responding to you, I thought the request for a bigger icon was reasonable. What I replied to was Bill Deville taking time to read the Apple HIG link that you posted and mock the wording, instead of just saying ok when we have time we’ll make a bigger icon.

At the time I thought Bill was representing Devon technologies official position on the topic and I found that reply snarky in the extreme and a very poor attitude for a software developer to have, no matter how great their program is.

It isn’t even that single detail, Patrick’s message was a great read that put a lot of this into perspective for me. What bothered me was the overall attitude Bill’s reply had, which was to mock something and not even consider why something as basic as a larger icon, might be important to some of Devon technologies users. You have the time to follow the link to the Apple HIG, read it, mock the wording and then post a snarky reply. Great, add up all the time you wasted doing that and you’re halfway done with just making a bigger icon.

If you read my reply as being to you CatOne, it wasn’t, sorry I’m not as articulate as some in this thread and didn’t use quotes.

@ Horace Greeley:

Hi, I didn’t mean to come off as ‘snarky’. I’m not one of the developers. I suspect that at some point DEVONtechnologies will use 512 X 512 icons.

One of my heroes of the early days of Apple is “Tog”, Bruce Tognazzini. He was a central figure in development of Apple’s Human Interface Guidelines.

Tog hasn’t been with Apple for some years, and has been critical both of Apple’s human interface decisions for OS X (including-over reliance on icons and a ‘flatlands’ approach), as well as their frequent and inconsistent departures from their own guidelines (albeit, sometimes with improved usability).

Tog advocates good design to increase the usability of applications. But he repeatedly cautions that bringing in a designer to ‘pretty up’ an application frequently makes the application less functional and less usable. Priority must go to quality assurance, to making the features work, and to get the bugs out (hence the public beta period of DEVONthink 2).

The complementary objectives of human interface design and software engineering should be functionality and usability.

Whether one agrees with Tog or not, some of his writings are worth reading.