Fog Creek Software
Discussion Board

Home office design for programmers?

Hi guys:

Inspired by Joel's office article but with much lower budget, I am turning a 2.6MX2.6M den (in a condo) into a programmer home office.

I googled for a while, but only get general ideas.

Is there any good resource for programmer home office design? (principle, rule , pictures etc.)

Some rules/principles so far in my mind are:

(1)All the outlets/USB ports should be the working table height.

(2)Computer/Server are organized on a light-metal frame.

(3)Only necessary wireless for security and pure signal reason


Please advice more.

Your ideas are very appreciated.

Monday, June 07, 2004

Make sure you have enough work to pay for all those perks.

Monday, June 07, 2004

You probably want somewhere to sit. I use a chair when I'm at the computer, and a sofa when I'm with paper-and-ink (I have a clipboard for the paper, so I don't need a table).

With the chair, you want something on which to place your monitor and keyboard. I don't see why you're saying "light-metal frame": I would consider that an implementation detail, not a requirement. I have a deep desk with a pull-out keyboard, so that the monitor is 130 cm from my eyes (I prefer that to its being too close).

You also want connectivity: internet and telephone. Wireless may be useful if you have a laptop and want to move your laptop around the house; otherwise, why wireless?

I don't care about having outlets/USB ports and table height because I don't often plug/unplug them. Also, more than "cozy" I like "well ventilated". I consequently have hardwood floor instead of carpet; and my chair is on rollers.

Drawers (with hanging files; could be filing cabinets) and bookcases are also handy, for reducing clutter. Noise (or music) can be an issue too, as well as temperature and lighting.

Christopher Wells
Monday, June 07, 2004

I wrote an article on that subject in my blog. Maybe I can add some photos to the article one of these days:


Karel Thönissen (
Monday, June 07, 2004

+ plenty of room under the desk for the fluffer

Monday, June 07, 2004

Ah, yes, the fluffer. That Fog creek feature Joel never mentions.

Monday, June 07, 2004

Dan Cederholm ( )covered some of the process he used to turn a walk in closet with window into a home office.  Of interest, he turned a large piece of particle board into a solid surface desk which added much detail interest, and used a dark color on the walls which made the room cozier (or claustrophobic depending on how you feel about such matters).

For my home office I designed and decorated it like a contemporary living room.  Comfortable couch, all electronics are hidden/secondary features.  Clean surfaces, dark woodwork and bright walls. 

Blue painters tape is your friend.  Get out some graph paper and draw your room - plan outlets and phones, etc - draw desks and couches etc to scale and play with the layout.  Then mask them out on the floor and really think if that will support your daily operations.

Monday, June 07, 2004

The biggest surprise to me in working out the office for my home-based business (which evolved over a number of years) was how much desk space I ended up needing before I felt like I had enough to be comfortable. Of course, Peopleware talks about this, but it was interesting to experience firsthand. I was only really satisfied once I had a total of three desks, each about 5' x 3'. (They're simple models without drawers, $100 or so apiece at Ikea). I also ended up needing a lot more bookshelf space than I had bargained for.

An item that I wouldn't want to live without is a wireless telephone headset. Clip it on my ear and it leaves both hands free for typing, or lets me get up and pace around the room. Also just bought a new phone with integrated Caller ID and a DND (do not disturb) feature that silences the ringer. Seems to help reduce unwanted distractions.

I've found good lighting to be crucial as well. For me that means abundant natural light (I'm fortunate to have a great view of the mountains from my desk) and also lots of good incandescent lighting to compensate for overcast days, late-night hacking sessions, etc.

I chose LCDs instead of CRTs -- we all know what the benefits are. Dual panels on my main dev system, of course.

Someone else mentioned the value of a good chair. I agree.

John C.
Monday, June 07, 2004

I agree with all the comments thus far -- lots of desk space, comfy chair, natural light (I have a little too much -- looking at getting some Hunter Douglas Silhouttes), cordless headset, etc.

One other thing is to make sure of is that you have enough electricity.  Not just outlets (your UPS & some power strips will get you enough of those) -- You need at least two separate circuits.  4 computers (even using LCD monitors) is about the limit for a 15-amp circuit as found in most US homes.  Add a coffee maker and/or laser printer, and you'll trip the breaker.  Make sure you tell your electrician to ensure both circuits are on the same phase, too.

Monday, June 07, 2004

OK, I get the multiple circuits.  Why do you care if they are on the same phase? 

Monday, June 07, 2004

if they aren't the same phase, you can get a 220V difference of potential between two "GND" pins across a cable. Or 220V across an Ethernet cable: in both cases, you really don't want that to happen.
Much, much less likely to happen in Europe.West, except if you're dealing with 230V made by picking individual phases of tri-phased industrial supply.  Typically won't happen unless the site is meant to receive heavy electric equipment.

It may sound obvious to point this out, but at least an el-cheapo UPS + data cable on each computer. Not necessarily to run hours and hours on battery power, but enough to cause a graceful hibernate, should the breaker trip...

(if you're using a VoIP phone, it may also be useful to UPS the whole chain from the external network interface to the phone handset, with perhaps a slightly beefier unit, at least matching the telco equipment's "running unplugged" capacity)

Cyrille Chépélov
Tuesday, June 08, 2004

Um, OK.  I'm be really surprised if we ran AC over any cables between any computers/peripherals/networks etc.  This is all low voltage DC.  PoE is 12V DC, para parts run at 5V, USB should be 5.25V.  Why should each component then care about the AC phase that is being rectified into DC?

O Canader
Tuesday, June 08, 2004

simple: you may have 5V DC between the two conductors of your typical wall wart's output, sure.

What may happen, though, is that the alleged "GND" output pin of that wall wart is actually connected to the lead input pin (yes, the one with (Locale.includes(USA)?120V:230V) AC).

Now, take a second wall wart, with the same flaw, plugged on another circuit, which happens to be on the opposite phase. So, again, the alleged "GND" pin is actually running at a high AC potential relative to the actual ground potential.

Power one device with each of these wall warts. Plug a data cable (with a GND conductor) between those devices. With a well designed insulation, you may be lucky. But if you have a less than good insulation within those devices, there is now a (Locale.includes(USA)?240V:460V) difference of potential between the two ends of that poor little data cable (before it melts).

OK, this is clearly a pretty bad failure mode, which requires to happen that several components are misdesigned or malfunctioning. But you don't want it to happen, do you?

Cyrille Chépélov
Wednesday, June 09, 2004

Take a look at:

In a typical US home, there are three wires from the power pole to your electrical meter: Hot1, Hot2, and Neutral (not the same as Ground, but close enough).  The voltage from Hot1 to Hot2 is 240v, which is what runs the high-usage devices in your house (clothes dryer, water heater, stove).  The same wires are also used in conjunction with the Neutral to produce two phases of 120v electricity.  These two "legs" are 180° out of phase with each other -- when one is at positive 120v, the other is at negative 120v.  These circuits are what run everything else in the house.

In the case of having two circuits for your computers, assume you have two computers, and their metal cases touch.  The cases are connected to the ground (aka neutral) wire.  But because one wire that runs to the breaker box is slightly longer than the other, there is a phase shift between the two circuits (speed of light in copper).  This causes a voltage difference in the two grounds.  In extreme cases you can raise a spark between the two computers and get shocked should you touch both of them at the same time.

Also -- remember that network cables, KVM cables, etc. contain grounding connections (for noise rejection).  These will carry the voltage difference, and again, in extreme cases, will burn out your network card, motherboard, etc.

Having both circuits on the same phase is more of a preventive measure than anything, but it's cheap to do, and the consequences of not doing it are so expensive that you really want to make sure your electrician understands your needs.

Also - double-check your electrician's work.  I've seen the Hot and Neutral wires reversed at the outlet with disasterous results.  Buy or rent an outlet tester.  Many UPSes have the ability to test for this too (not the cheap ones, though).

Wednesday, June 09, 2004

But computers should be grounded on the ground wire not the neutral wire -- this is why we have 3 prong plugs. 

Wednesday, June 09, 2004

The neutral is tied to ground inside the breakerbox (plus in the meterbox, plus on the pole outside the house, plus at the substation, etc).  But like I said earlier, the individual wires don't always have the same electrical lengths.

When I was in the Air Force, one of the yearly maintenance inspections on our equipment was to verify that the grounding was good.  While doing this, I discovered that there was a 20v difference between the neutral and the ground.  It turns out that the wire to tie the two together was missing in the breakerbox, and the difference in electrical lengths (from the equipment to where ever they got tied together -- probably at the building's transformer) caused this 20v difference.  While this wasn't a safety hazard, it shouldn't have ocurred, and was corrected by the base civil engineering team.

Wednesday, June 09, 2004

*  Recent Topics

*  Fog Creek Home