Hacker News Reader

Comment by jchendy | Hacker News Reader

IBM actually seems to be taking design very seriously. They're hiring 1000 designers, and the design program has significant support from the CEO. More background here: http://www.nytimes.com/2015/11/15/business/ibms-design-cente...

View on HN

Is there a design equivalent to the mythical man-month in software engineering?

Interesting question. Theoretically, no, because ideas are a dime a dozen, and there's no "one true way" to solve a design problem. But I think what you'd run into over time, which may be parallel to the man-month myth, is a higher-order approximation of groupthink. Meaning - "Let's apply this interaction pattern to solve this class of problem because we've done it before," repeated ad-nauseam. This will actually result in that parallel of the man-month myth - "solutions" (note the quotes) will be churned out at an increasing velocity, like a giant rubber stamp factory. The big difference is obviously ideas are much easier to create and document than software is to build. I don't think this is a result of the User Centered Design process, however. I think it's a result of the oversimplification of the interaction design sub-process, coupled with the broad lack of interest in the business improvement part of the process.

The problem with design studios this large, and operating on this consultatory premise, is that they will never get to influence the business "why" of solving a problem - they will always be handed a brief, and will always have their solution bounded by decisions made prior to their involvement. They are compelled to do work on what someone else has determined to be the problem, not to uncover the problem themselves and work outwardly from there.

You actually see this all the time in finance - there are entire companies built around the idea of churning out datagrid UIs to banks that don't ever stop to consider the real underlying conditions of why a process is the way it is, or have the opportunity to change it. It just results in a lot of same-same being generated over and over again, and it's easy money for consulting groups to get into the biz - just buy a small (or several) design shop and start billing yourself as being in tune with "UX." Deloitte, KPMG, MarkIt On Demand, EY, PWC, Accenture, EMC, and a bunch of other places are doing it just that way.

Yes -- It's Fred Brooks' own "The Design of Design:" http://www.amazon.com/The-Design-Essays-Computer-Scientist/d...

If you assign 1,000 designers to 1,000 different products then it avoids scaling problems.

Yep, I'm on an engineering team at IBM and this is basically what's happening. In my (admittedly limited) experience, I can only think of one thing that had multiple designers working on it, and it was basically a marketing website.

There are design guidelines and even some base templates and sass files and such to keep things looking somewhat similar.

I'm a recent hire at IBM Design and leading a design team.

nfriendly, your experience is what it's like working with a non-IBM Design team. The old way is to have a single designer on something and that person has little to no authority to be the users through the design process. That's all changing now.

From my understanding, there have only been a couple of hundred products (in the 3 years since IBM Design was created) that have been updated using IBM Design Thinking. There is actually a dedicated design team (min 5 designers) working on each product that is getting attention at that moment in time.

But there are over 5000 products that IBM sells, so it's going to take a while to get to them all (also assuming some of these will be consolidated as the design cycles address the user's needs). There will be a lot of time that passes before all the engineers at IBM (with it's 400K+ employees) get to have a dedicated design team on their product.

Having been a number of startups and advertising agencies before this, the "design culture" is strong here with a lot of smart folks fighting to turn around a very very big ship.

However, then you wind up with 1,000 different products with 1,000 different design languages, if the designs are even thought through well enough to be considered having a design language.

Not necessarily. One person defines the design language and the other 1,000 apply it to different products.

1,000 people with bad taste do not make a Sir Jonathan Ive.

I get the feeling adding designers to a project has the same startup costs as hiring new programs with the results of crappy sorta conforming to the theme design being just as noticeable.

I believe it's called Design by Committee.

1000 monkeys throwing feces at the wall...

Very good question. Shame no replies

Of course it's very important to have a beautiful Z System mainframes. I like the effort they are putting into it, but do they customers care (even for the software)?

Design is more than how something looks. To reframe your question, you might say "but do their customers care if the stuff they buy works?" To which I hope they do care.

I definitely agree with your point, so I apologize for nit picking.

IMO, "does it work?" is a QA question. The design question would be "does it do anything useful?"

FWIW Since IBM bought Silverpop the product has had a gigantic design (and functionality) upgrade. It is very noticeable and has made the product feel far less painful to use.

IBM dont have a resources or Engineering problem. I dont think they lack talents either.

And their biggest problem isn't design.

It is their culture, and I would argue a lack of focus and direction.

Reply on news.ycombinator.com