|
Monitor size and developer performance
I have to write a business justification to upgrade our monitors from 17’’ to larger monitors. The idea is to convince the CEO that our team (4 developers working on a data visualization and analysis desktop application) will perform better and will deliver faster if we get decent 19’’ or larger (preferably TFT) monitors.
This might seem a no brainier but the team has just started and someone though that 17’’ was the norm and bought new monitors of that size for everyone :(.
Are there any studies out there that relate larger monitors with better developer performance? Any examples you can tell me about?
I Googled around but I found mostly ergonomics related stuff, which is relevant but I think will not be enough to convince the guy.
JSD
JSD
Saturday, April 17, 2004
Law Of Increasing Returns To Scale
=========================
Developers are unsociable, aloof and restless. They have no lives. They are sexually frustrated. An increase in the scale of monitor dimensions will tend to improve the developer performance*. With a large monitor, the restless developers will see larger porn images, larger busts, large cocks etc. This will make them happier. They will perform better in bed.
----------------------------------------------------------------
Footnotes:
*Performance = Developer Libido.
Keynes
Saturday, April 17, 2004
I don't have useful stats for you, but if the person in charge of money-spending at your company doesn't want to "waste" the current investment they just made in 17" monitors, perhaps they might feel better about adding a 17" monitor to each workstation as opposed to getting rid of the 17" monitors they just paid for?
Just a possible suggestion to perhaps make the purchase more palatable...
John Rose
Saturday, April 17, 2004
Larger monitors mean happier developers.
Larger monitors mean higher resolution.
Larger monitors give you more workspace.
Larger monitors mean that you have lesser windows to close, lesser windows to scroll, lesser windows to move around and more windows to be viewed at the same time.
Larger (and FLAT) monitors are ergonomically better for the developers. Less sore eyes mean they can work longer, more effectively and can be more productive overall.
Green Pajamas
Saturday, April 17, 2004
Thanks for the suggestions.
Adding another monitor to the existing ones is a good option. But I guess that would mean buying special graphics cards (there's a very nice dual head Matrox out there) and that would have to be money justified all the same.
Developer happiness is not very high in their list so I’d rather not use it. Actually, I am making very clear that all of us have huge monitors and several computers at home (which is true). I find very patronizing the prevailing feeling that developers have to be given toys to remain in the job. What I need to demonstrate (preferably backed with numbers) is that the job requires larger monitors.
JSD
JSD
Saturday, April 17, 2004
There have been a billion threads about this before, but you can get a nice dual-monitor card for about $31. Check NewEgg for "RadeonVE".
The Matrox cards would have slightly sharper output, but the ATIs are probably 95% as good for, what... 15% of the money? Just FYI.
John Rose
Saturday, April 17, 2004
Any manager should know that buying the monitors will increase productivity.
Why? Because the developers say they will be more productive; so if they get the new monitors they will prove they were right by working harder, at least for a while.
I'm sure psychologists have a name for that behavior. Joel might even know the term, since he seems to hire programmers who place a high value on spiffy offices (which of course will make them more productive. Why? because...)
Tom H
Saturday, April 17, 2004
FYI, a 17" and 19" LCD have the same native resolution: 1280x1024. The 17" will have approximately 96 DPI, the optimal for Windows; the 19" will have a lower DPI rating, and the pixels will probably look too large (they do to me, anyway, especially when you compare them to laptops that have DPI ratings between 120 and 150!).
You need to step up to 20" to get 1600x1200. An excellent option for a 20" LCD monitor is Dell's FP2001, which can usually be had for ~ $900 each. A 20-21" CRT is going to be significantly less expensive, and may be easier to justify since you don't have an understanding boss.
Brad Wilson (dotnetguy.techieswithcats.com)
Saturday, April 17, 2004
Just time someone on a single 17" for an hour. Count the time spent resizing windows and the time spent switching from one window to another, and remember to add a few seconds for the irritation. Then extrapolate the seconds in an hour to minutes in a day to minutes in a week to hours in a year and multiply by three for the three year depreciation cycle.
Whether you are better off with two monitors than with one large one depends. I would say that for all but graphic artists the answer is yes, since the real hassle is switchng from one window to another (say the help window to the IDE). As you have already got 17" monitors, then you are talking the cost of a 15" TFT and a second video card.
Stephen Jones
Saturday, April 17, 2004
Stephen, I am going to try that (though I am not very sure how :-) to come up with a number.
After some more research my plan is to convince them to buy 19’’ CRTs plus dual head cards (TFTs will most probably be dismissed as too expensive).
We already have one 19’’ and one spare 17’’
We would need to buy:
3 X 19’’: £150 X 3 = £450
4 X Matrox G550: £80 X 3 = £240
Total : £690
At some £20/hour per developer the company spends about £100/hour for the team.
If each developer wastes 10% of the time switching apps and that value would come down to 5% then that would be a £5 savings per hour for the team. The setup would pay for itself in 690/5 = 138 hours -> roughly 20 days at 7 hours per day.
Even if it takes three times that amount of time the balance is zero after 3 calendar months. The problem is proving that larger monitors do make a difference. Hard data or even a case study would make my case so much stronger. I’ll try and do some measurements myself.
P.S. These numbers are ridiculously low(!) but that’s life in IT these days:)
JSD
JSD
Saturday, April 17, 2004
"Developer happiness is not very high in their list so I’d rather not use it."
Nice. Nothing like a company that doesn't care about morale. How about:
- If the developers have better machines at home then they'll be out the door at 5pm. If they have better machines at work, they'll stay all night. (gross generalization). And, bizarre as it may seem, for a lot of devs it's not so much what code they're working on as the system. :)
- If the developers have better machines they're less likely to go work somewhere that will give them better machines.
- If you're interviewing a developer, he's more likely to accept the offer if there's a 20" flat panel on every desk.
- If you do contract work and show clients around the office, they're going to be much more convinced of your capabilities if there are two 20" flat panels on every desk.
Regarding numbers - a 17" CRT runs best at
1280x1024= 1,310,720 pixels
1600x1200= 1,920,000 pixels
= 50% more screen real estate, which means things like having two documents open at once, or a document and a code window, or the IDE fully expanded with tool windows open.
(BTW, you can get a decent 21" CRT for US$300, and I run mine at 1800x1440 = 2,592,000 = twice the screen real estate)
Since my big laptop was in the shop I spent four weeks working on my tablet at 1024x768 and I can tell you I was honestly less productive (and miserable)
Philo
Philo
Saturday, April 17, 2004
Oh, and if you get those dualhead cards, then the 17" monitors won't go to waste. ;-)
By the way, if management does this, it will be an enormous morale boost - to come back and say "Hey, we made a mistake by not asking you, so here's our way of making it up" - you can't buy loyalty, but this may rent it for a while.
Philo
Philo
Saturday, April 17, 2004
Philo,
Don't misunderstand me, the company is not bad at all, its just a matter of saving as much as possible until sales start in earnest (small company, just starting, not a lot of money lying around...).
Having said that, the screen estate and the two documents side by side angles are good and easy to demonstrate.
I have a 21'' at home and I just tried the 1800X1440 setting and god it hurts! (had to go down to 70Hz to do that). And my code looks like a newspaper column too.
JSD
JSD
Saturday, April 17, 2004
I have a colleague who judges (or 'whack-a-moles') how a company treats its developers by looking at their monitors: if employees have small monitors then he won't accept a job there.
Christopher Wells
Saturday, April 17, 2004
p.s. if cost is an issue and you have a spare PCI slot Windows does multi-monitors just fine with two or more video cards. Before I got my dual card I had an AGP and an el-cheapo PCI card (you're not gaming or doing CAD). Worked great.
MR
Saturday, April 17, 2004
JSD - yeah, I've always been a real estate hog - I ran 15" at 1024x768, 17" at 1280x1024 or 1600x1200, and this beastie at 1800x1440.
I run my 15" laptop at 1600x1200. :)
Philo
Philo
Saturday, April 17, 2004
If you can't justify a 21", at least shoot for 2 19", if you can't get justify a 19", at least get 2 17". In the end it's all about the pixel count without the squinting.
I can't imagine the additional radiation will be any good to your health though.
Li-fan Chen
Saturday, April 17, 2004
On a 17" CRT running at 1600x1200 is pointless. The pixels become smaller than the dot-pitch. You don't actually get more information onto the screen as opposed to going into settings and turning down your font size.
I've a 19% CRT here; it's adjusted so that the pixels line up with the vertical stripes. This worked out to be 1280x1024 with a little black at the edges. I can now run text smaller than I used to, because the edges are sharp and well defined.
Ham Fisted
Saturday, April 17, 2004
All of my developers have dual 20" LCD monitors with 1600x1200 resolution. I'm starting to look into larger higher-resolution displays, but probably won't do anything about it this year.
The business justifications that I see for the large monitors all revolve around being able to view more information at the same time, both documentation and code. A 20" 1600x1200 display allows reading a full letter-size page, plus having additional space for navigation information. A 20" 1600x1200 display also allows having two readable pages of code side-by-side. Smaller displays won't do for this. Even rotating a display 90 degrees won't do, as you lose the ability to do side-by-side display. Most coding involves making use of existing libraries, so being able to view both your code and the library definition at the same time is extremely helpful.
A large amount of the documentation we use is in Adobe PDF format, and being able to view full pages plus navigation at a same time makes finding information much faster. We've eliminated printing.
For coding, the more you can see on the screen at one time, the better. Many bugs are the result of incompatibilities between pieces of code. Being able to view multiple full pages side-by-side simplifies writing and reviewing code while at the same time viewing the definitions of the functions being used.
As an example, a few days ago I was working on some code that involved two specifications from standards organizations (full-page PDF), code supplied by a chip vendor for reference, and new code I was writing myself. This is easy when you have enough screen real-estate.
Our bug rate went way down when we switched to dual 21" CRT monitors. The 20" LCD screens are certainly more pleasant to work with than CRTs were.
Dan Brown
Sunday, April 18, 2004
Work out how much time you'll spend preparing this business case, and how much time will be spent discussing the issue in absolute hours.
Figure out how much that time will cost the company. The chance is good that time will cost more than getting decent larger LCD displays for every developer. (Possibly even more than high-end LCDs like the Apple Cinema Display HD.)
Write out one paragraph saying this as your business case.
Chris Hanson
Sunday, April 18, 2004
Here are a couple of Microsoft Research documents with a bit of quantitative data:
http://research.microsoft.com/displayArticle.aspx?id=433
http://research.microsoft.com/vibe/Interact2003Productivityl.pdf (PDF)
Or maybe you could convince the approver to try using a dual-monitor system himself for a few days. Once you start, the benefits become so obvious that it's painful to consider going back. Maybe he'll get the idea...
John C.
Sunday, April 18, 2004
Very good Idea John... I was thinking of suggesting a test case where JSD for one week pin down how much he gets done and then for an other week where one of his collegues is on vacation borrows the monitor (and buys a video card if needed) and stats the stuff he gets done in this week.
Peter Monsson
Sunday, April 18, 2004
you can even go with dual 17" monitors for some people and larger monitors for others--take your existing stock & move it to half the people, and buy the half-number of larger monitors. i much prefer dual 17" over single 21".
you will have to buy new video cards, as someone mentioned you can get any old single-head card to add to whatever you already have. but do try to match frequency ratings, i have to run one monitor at something like 79.8 kHz and the other at 80 kHz, which leads to a stripe moving through the display as the two monitors aren't quite in sync.
mb
Sunday, April 18, 2004
Wow John! That's exactly what I need!
Naturally the highest gains in productivity happen with 21’’ monitors, so I will present the case for 21’’ first and scale back to 19’’ if the first one is refused. Also the gains can be higher than I expected, going as high as 50% (for cutting and pasting). On average though the increase is on the 9%, which is high enough to repay the monitors very soon.
One thing worries me though; I have a 21’’ next to 15’’ at home that I only use when I need to debug UI stuff. The problem is, as mb pointed out, the differences in frequency (the graphics cards are also different which makes it even worst) – it is an instant eye killer. That’s why I am shying away from proposing one additional card instead of dual cards.
Following the references of the article above (the second link provided by John) I found a study called “Effects of Screen Size, Screen Resolution, and Display Rate on Computer-Based Test Performance” which adds interesting reading:
From the abstract:
“Computer-based tests administered in established commercial testing centers typically have used monitors of uniform size running at a set resolution. Web-based delivery of tests promises to greatly expand access, but at the price of less standardization in equipment. The current study evaluated the effects of variations in screen size, resolution, and presentation delay on verbal and mathematics scores in a sample of 357 college-bound high school juniors. The students were randomly assigned to one of six experimental conditions—three screen display conditions crossed with two presentation rate conditions. The three display conditions were: 17-inch monitor set to a resolution of 1024 x 768, 17-inch monitor set to a resolution of 640 x 480, and a 15-inch monitor set to a resolution of 640 x 480. Items were presented either with no delay or with a five-second delay between questions (to emulate a slow Internet connection). No significant effects on math scores were found. Verbal scores were higher, by about a quarter of a standard deviation (28 points on the SAT scale), with the high-resolution display.”
From the discussion:
“A major difference between the high and low resolution presentations was the substantially greater need to scroll through reading passages in the low resolution condition. Furthermore, scrolling was identified by the majority of the participants as something that interfered with taking the test.”
http://www.ets.org/research/dload/AERA2002-bri2.pdf
The article should also be of interest for anyone designing CBT stuff.
Finally, if all fails, I’ll definitely try to prove it using Peter’s idea and make the case again and again.
Cheers
JSD
Sunday, April 18, 2004
FWIW, I have a 21" CRT and a 19" LCD, which I run at different resolutions and frequencies, and I don't have a problem. Two caveats:
1) They're significantly offset, so that generally one is in peripheral vision, instead of them both being in my focus area.
2) My right eye is 20/40, my left eye is 20/400. Every optometrist I've been to says this is very unusual. My experience is that I'm not bothered by "different views" like some people seem to be (I would often just wear one contact lens if I was in a hurry).
In other words, mixing CRT and LCD may solve the different scan rate issue, or I may be an anomaly.
Hm. I guess this post added no value whatsoever. Sorry. :)
Philo
Philo
Sunday, April 18, 2004
the big problem is that the CRTs are next to eachother--an LCD doesn't really have scan gun like a CRT. in my current setup, the best match is 75hz and 75.2hz. there's a faint stripe running through the monitors which takes about 24 seconds for a full vertical scan--not particularly irritating, the previous offset was about 5 seconds.
btw, i run each monitor at a different resolution (1024x768 and 1152x864 right now), so i can have easier reading on one and more text on the other.
mb
Monday, April 19, 2004
You think 17" is small? I am trying to get our company to upgrade to 17" falt screens. Currently all junior employees have a monitor twisted to the side of their desk so they have to contort their entire body to use it.
I like the two monitor solution, it is great for debugging.
Aussie Chick
Tuesday, April 20, 2004
 Recent Topics
Fog Creek Home
| 
