As a former trader, I can say that Bloomberg's UI is, while not perfect, in general quite good. You can do almost everything straight from the keyboard; once you get used to it, it's quite fast. In about 30 seconds, I can calculate the yield on a mortgage-backed security, see what the Treasury market is doing today, skim a list of top headlines, and show a graph of the stock market since today's open. Try doing that with Yahoo Finance!
In that sense, it's very much in the spirit of any other command line environment. But apart from any UI issues, people continue to use it because of network effects. It's got a great messaging system that everyone uses, and it's got a common set of tools shared across the entire financial system.
Old habits die hard, too. Even though I'm no longer trading, I still have an old Bloomberg keyboard on my desk.
Replace "can" with "should", and I'll happily agree with you.
I'm a vim fan, personally. The keyboard is ideal for high-speed, high-skill interaction. Most GUI-centric applications fail completely in this area, where a lot of old applications had a lot of focus on it. Both is probably the best, but that's hard work and doesn't always fit with the GUI without it being cluttered for a "clean" interface.
Definitely agree. A lot of people use their Bloomberg Terminals completely by keyboard, whenever they even so much as change the tab ordering of fields on a function people complain. A lot of people have the layouts/ordering memorized.
Think about how much people complain when Facebook change their UI. Now imagine that if when they changed their layout you weren't able to do your job because it took you twice as long to find some information or if you lost a lot of money because you missed a pricing alert. Users are opposed to change for many reasons other than complexity for complexities sake.
If your company is willing to shell out $2k/month to give you a Bloomberg you've got other stuff you can brag about far more than the fact you can use a complex piece of software.
But I imagine the real reason they haven't done anything is simply because it's not worth it. Investing a huge amount of money (we're talking in the high tens or low hundreds of millions here) to revamp your GUI without a clear return on investment would be idiotic. They have been tackling the low hanging fruits though (the new graphs, font, etc.).
This looks like yet another case of designers not "getting" the needs of anyone who wants to use the software for any length of time.
I took a gander at the comparison on http://www.portfolio.com/interactive-features/2007/06/termin... , and the whole article just annoys me. If they could focus less on "oh dear, this isn't going to win any design award" and more on making the software more efficient to use, I am confident they would not have faced the wholesale rejection that they complain about.
The current design is easy on the eyes. A black background with bright text is easier on the eyes than a glaring white flashlight with some thin black text on it. (I wonder if the misconception that dark background is bad stems from print, where it is a legitimate complaint?) Fun test: try using each of black-on-white and white-on-black at 3AM with bloodshot eyes and see which one gives you a pounding headache and which one allows you to focus on your work.
The current design is compact, and uses nice color-contrast techniques to group regions, while the "improved design" wastes a ton of pixels for spacing between the regions instead. Less eye-travel, less scrolling, less next-page-ing makes for more efficient (and therefore more relaxed) use.
Emacs or vi won't win any design awards, yet it is the editor of choice for a lot of hackers, and not just because of machismo.
This is because (A) there is more to design and usability than visual design, and (B) some systems design should be optimized for the expert user, not the general user.
I don't know if I can agree with the stick shift argument. There are benefits over an automatic transmission, and you can choose the same model of car with either transmission types. I'd equate the differences to having a settings page with defaults and then advanced user options. I like a better design than not, but I think the traders use bloomberg terminal largely for the data that it displays. I don't think that is is going to change much between the current and potentially redesigned versions.
Agreed. As far as I'm concerned, a manual transmission and an automatic transmission are completely different animals altogether. In my mind, his whole argument fell thru once the statement was made. It's like comparing a text-based shell with a GUI interface. There are trade-offs to be considered and neither of them are "better" than the other, they just have different strengths and weaknesses.
Since my transmission is stick shift, I'm not worried about Toyota floor mats, accelerator or brake pedals, or software bugs. I've never had a problem, and I would reflexively respond in a safe way were something to occur.
Pride in their ability and job security are two reasons why people may prefer such user interfaces. Some complex user interfaces may also give people the perception of having more control.
These statements are incorrect. Command line interfaces (black background, in my case green text instead of the article's yellow) often give more control and are more usable than "simple" GUI ones. Especially for complex systems, it becomes unwieldy when a GUI interface tries to provide graphical settings for every supported feature. There are a bajillion tabs and windows, and while the UI designer though of several brilliant one word descriptions for each settings section/tab, of course they're not completely accurate for the setting which you're looking, and you end up flipping through way to many tabs/windows.
I once worked for a company where the Marketing department believing anything that was not dark text on a white or off-white background was hard to read. Amazing how a couple individuals could off handedly discard an entire rainbow based on a single mis-guided principle: http://www.smashingmagazine.com/2010/01/28/color-theory-for-...
The fascinating thing about graphic designers is, if you use Photoshop then what do you get, a tool palette, in stark black and white, full of cryptic symbols. In other words, they tacitly acknowledge that for a busy professional, learning a "difficult" interface is fully worthwhile. Yet they continue to peddle "design" as if this isn't the case...
In the case of Photoshop I think it's more that an established system with a large userbase can't significantly change its UI without significant negative feedback (see Office Ribbon).
The pallet interface is also what we're used to. Most people, were they to set out to build a pixel pusher app, would start with a basic pallet UI. Unless of course they were specifically aiming to get away from that.
Also, Photoshop is a perfect example of "The more painful the UI is, the more satisfied these users are".
I guess it's just preference, but you can see this on bloomberg.com as well. I find it blinding when switching from a page with a white background (as most have). When using the terminal itself, its not as bad for some reason, but then again I don't use it on a daily basis.
BBG's GUI is not pretty, but the real reason people use it, and pay a hefty fee at that, is because it is full of things people need, and that's an understatement.
It's not just UIs. I was in the closed beta for a computer game (friend was a dev), and one of the goals of the developers was to lower the barriers to entry for new players to get into the genre. Nearly universally, the self-described 'hardcore' players decried the changes as dumbing down the game and removing the skill barriers between new and old players. In fact, some of those changes were UI upgrades, like making the item shop easier to navigate. While that one was more generally accepted, there was still a vocal minority that thought that noobs being able to easily purchase the items they had earned would lower the competitive aspect of the game.
While there's often an initial loud and forceful resistance to change, it dies down as people get used to the new system, except for a few die-hards. In the example above, it took about two months for the community to come around and realize that removing some obvious pain points let the players focus more on strategy and whatnot. Of course, it helps that the new system is clearly superior to the old system. Microsoft Office's Ribbon, on the other hand, is not clearly superior, so people still bitch about it years after it was introduced. I happen to think it's superior in theory, but the implementation was botched a bit (menu names and organization not intuitive), so it's not a clear upgrade.
I browsed up to the uxmag.com homepage because it sounded interesting. Is it me or is that an absolutely terrible UI/UX design? I took one look at it and felt like giving up.
The Bloomberg terminal is the perfect example of a lock-in effect reinforced by the powerful conservative tendancies of the financial ecosystem and its permanent need to fake complexity.
I've noticed relentless attacks on wall street portraying them as dumb people who do simple work but are just ruthless and greedy as hell. Of course they're greedy risk takers, and sometimes that causes problems. However, financial models can get fairly complex and require intelligence to navigate and utilize.
I was listening to NPR post-Haiti earthquake, and they were interviewing this woman who was basically borrowing money, going to the Dominican Republic to buy goods, giving them to stores in Port Au Prince, and then collecting the money two weeks later. She also charged the shopkeepers interest. When the earthquake hit, all of her goods were destroyed in the shops, so there was no way she could repay her debts.
Anyway, the radio hosts kept harping on the fact that she had only a middle school education, and was doing "extremely sophisticated math, the type that people on wall street get paid millions for". Now, I very much respect what that woman (and she was working off her debt post earthquake), but I thought it was incorrect and misleading to bring in wall street.
There are a lot of other examples, these are just two.
Shouldn't this entry link to one of the original articles instead of to some random, er, symbiont? Sorry if I'm missing something here; I'm the new kid.
Yes, but no reason to fling insults around. Linking to articles you enjoyed and commenting on them is not some kind of despicable act. It’s, in fact, quite normal.
I didn't mean to be insulting, it's just that I don't see how this particular instance of blurb-cum-comment page adds a lot of value. I'd have thought it would have been more helpful and polite to skip the intro and link to the full article straight away... but then again I'm new here, which is why I asked.
Does HN encourage this kind of redirection so that the blurb linked to may serve as an executive summary or tl;dr or whatever you want to call it? Does HN encourage this kind of redirection to elicit submissions from people with blogs to draw attention to?
In general the original source should be linked. Certainly in this case.
No reason, though, to call the author of the submitted article a “symbiont”. You could do that if the author of the submitted article is also the submitter (not the case here), but simply submitting the original article would probably be more appropriate.
I'm quite annoyed when I get a GUI with no keyboard shortcuts; it just doesn't occur to some developers that being able to use the keyboard is what allows advanced users to become very fast at using the software. Full on command line capability for very advanced users is great, but rare.
I remember a UI I was given to test for controlling a custom video display; the operator needed to be able to update things very quickly, but there were zero keyboard shortcuts defined, so everything had to be done by move-click-type move-click-type. Keyboard shortcuts hadn't even occurred to the developer as being important, despite the speed requirement.
I don't know anything about Bloomberg, and I'm willing to bet that as a niche product aimed at high trained people, it might not be too bad.
But I agree in principle with this article. You can see it sometimes in the Linux/FLOSS world. People decry anything that means the average user can use it.
"But you should know how to set up networking on the command line!"
People like that, they like jumping through hoops. They think sensible defaults are a bad thing. They think you should have to read a 1,000 page manual before doing something simple.
It's actually not about jumping through hoops, it's simple math. Doing a command in a terminal is O(1). Everything is at your fingertips immediately and can be combined efficiently with everything else. Doing something in a GUI is O(n), you have to dig through menus and tabs to get at what you want, and the result is in a form that can't be recombined or filtered or modified at will.
The trade-off is that the command line takes a lot of getting used to in terms of memorizing commands etc. If you don't know any Unix commands, the terminal is almost completely useless unlike a GUI. For beginners, a GUI is a good way to explore and get a feeling for how to do things because GUIs explain themselves.
In that sense, it's very much in the spirit of any other command line environment. But apart from any UI issues, people continue to use it because of network effects. It's got a great messaging system that everyone uses, and it's got a common set of tools shared across the entire financial system.
Old habits die hard, too. Even though I'm no longer trading, I still have an old Bloomberg keyboard on my desk.