Hacker News new | past | comments | ask | show | jobs | submit login

As I recall, Microsoft designed X-shaped buildings with shared spaces towards the middle.

Every developer has a private office with "a door that shuts". They all have a view that doesn't look into another office. When they need to work alone, they can work alone.

Each office is large enough that a second person can come in and work alongside them, whiteboard together etc.

Social areas are placed towards the centre of the X. Offices are on the outer. This controls the spread of noise from the social areas.

These buildings were inspired by IBM's Sillicon Valley Lab (née Santa Teresa Lab), which itself was inspired by studies of programmer productivity which found that private offices improved performance markedly over cubicles and open-plan offices.

There's a discussion in Peopleware.




Where's the evidence for these claims though? I know Joel Spolsky makes the argument about "A door that shuts" but it's purely anecdotal. http://www.joelonsoftware.com/items/2008/12/29.html


It seems that the X-shape exists as described, though not for all buildings:

http://www.microsoft.com/presspass/ImageGallery/ImageDetails...

Some buildings aren't strictly cross-shaped but still use right-angled wings for a similar effect:

http://www.microsoft.com/presspass/ImageGallery/ImageDetails...

Compare STL/SVL:

http://www.nfbcal.org/~anordley/IBM/Air/SVL/

IBM published a paper on STL/SVL in 1978; I haven't gotten around to piercing the IEEE paywall to read it yet: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumb...


F%!#, what a ripoff. IEEE want $31.00 for that paper.. and sadly, it's the same price whether you're an IEEE member or not. So much for the idea that one of the advantage of being an IEEE member is that you get better pricing on papers from journals you don't already subscribe to.

Even sadder, IBM used to publish their journal archives online, available to anyone, for free. IBM Systems Journal wasn't always behind some shitty IEEE paywall.

Yet sadder still, the price to subscribe to IBM Systems Journal through IEEE? $1,400 USD for a one year sub.

Frack... I now remember why I'm sometimes ashamed to admit to being a member of IEEE. Guess it's time to get move involved in the elections and leadership stuff... sigh


I agree. It's extremely annoying how IEEE carefully segment their library access to maximise revenue instead of member value.


Shit, I even got an insurance (car, I think) offer co-branded from my IEEE membership. I've not renewed.


The study in Peopleware was individual coders working on well-defined toy projects. Whatever evidence taken from that may or may not translate to your real-world development organization.


You're right software engineering literature is not as strong as, say, physics -- where you can relatively easily strip out confounding factors. Or medicine, where you can get funding for very large samples. We have to work with the limits of the subject and the research funding.

However it seems that there is a growing body of supporting literature coming from other fields that open-plan is usually a decision with unseen costs. The linked article is an example.


of course not it saves building services budget and they dont care a fig for the effect on the business overall.

Some director saved 10% on deskspace costs and got his/her bonus but screwed the productivity of all the staff by 10/15%




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: