NSF Wants To Know How Much Software Really Costs
eldavojohn writes "It's no secret that the actual cost of software is very complicated. Sure, the companies that write software are spending money on it, but when that software is released, it doesn't stop costing money. You can probably think of a number of relatively tiny things that add up — especially if you're a system administrator — like the man-hours spent patching software to avoid a nasty infection spreading quickly. The bigger debt is that old piece of software you paid a bunch of money for back in 1998 that you're critically dependent on, but it has no support and hasn't been updated in years due to any number of reasons. Well, the National Science Foundation paid Gartner almost half a million dollars to find out what it truly costs to bring an organization to a fully supported environment. According to Gartner, this hidden liability or 'IT debt' is at $500 billion worldwide right now, and in five years it will be at $1 trillion. Along similar lines, a company called Cast that makes software quality tools reported that your average business application comes with a million in IT debt (PDF). And if that's not misapplied enough for you, they estimate that the debt is $2.82 per line of code in the application and also that it's on average higher in the government sector."
It could cost $5,- per game and people would still make big profits. Illegal copying drives the price up, however.
-- Cheers!
That doesn't sound too bad.
Software metrics are always a huge load of bullshit. It doesn't matter which "methodology" you choose when making such "measurements". Any time that something can be distilled down to "cost-per-line" or "errors-per-line", you can know right away that it's total nonsense.
The common estimation formula where I work is $10,000 per verb. Considering all the bad / inconsistent requirements we get that sounds about right.
I think we should be able to do better, but I'm too exhausted by my job to truly care. Oh well... off to work.
I know companies that don't bother figuring out the 'hidden' cost of keeping their workstations or servers up to date. Then one day they realize they need to upgrade 30+ system all at once for some new piece of software they want. When they can't budget/manage/understand something as straightforward as hardware maintenance and upkeep, how are they going to understand something less physical like software 'debt' or whatever they are labeling it now.
I will shred my adversaries. Pull their eyes out just enough to turn them towards their mewing, mutilated faces. Illyria
No, unauthorized copies of software are not driving prices up. Prove that the users of those unauthorized copies would have paid for an authorized copy, and you might have a case.
Why must we repeat these arguments over and over again?
Nice joke. If I say something completely bullocks I get one point from you and -4487564 points from other moderators. If I say something insightful or whatever, adding valuable content to the thread I get my positive modpoints from everybody and don't need your pityful one point ;)
And bloody hell, now I have to wait 5 minutes before clicking 'Submit' because of /.'s annoying posting rules. That annoys me no end.
-- Cheers!
Next, the automakers will get creative and figure out the national "automotive debt" that details the long-overdue cost to update everyone's wheels to the current standards. Of course, once they do, the "debt" starts rolling again.
And don't forget the national "old charcoal grill debt".
As for IT, if everyone upgraded to the latest software that would make life easier for hackers, no? They wouldn't need to waste cycles studying the vulnerabilities of 10-15 year old applications each used by a widely scattered customer base.
The NSF wants to know something about the computer industry and they ask Gartner? Gartner, the company that advocated OS/2 and I-CASE?
Hehe. We had one of those IT department brainstorming sessions once (I was in research at the time) and they were talking about this shiny new platform that they were going to roll out, I simply asked what the cost was. They threw out some figures about how they priced it an it would cost X dollars to implement over Y years. So, I asked "does that include the cost of decommission?" and got blank stares all around... The notion that you estimate the cost of getting out of abandoning / migrating away from a product never occurred to them! Products tend to not be all that flexible, they change over time, and business needs and processes often diverge from the product or a better product comes along -- we have fairly good ideas on what the platform turn-over is going to look like, how open various platforms are, etc. We can estimate the CoD with some accuracy. So why don't we? We're still buying into products that are readily identified as "dead-ends" and screwed when they are no longer supported, needs change, etc.
It's just a number. Calculating it is complicated.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
Obviously software cost depends on what you measure it in. For example Linux kernel is estimated to cost near 1.4 billion US dollars (at the bottom), but IF you measure this in chickens.... it could cost 35,008,752.2 chickens.
In ounces of gold it would be around 1,040,041.6 ounces. In DOW it would cost approximately 127,186
It is also possible to estimate its cost in terms of Libraries of Congress, man years and many such wonderful things, however note that many Keynesians say that gold has no value but what is 'speculated' to be value while they do not see the same thing about their cherished and printed fiat, so then we could argue that Linux kernel is worth nothing if 1,040,041 ounces of gold priced at current levels in USD are worth nothing.
It's all a matter of point of view.
You can't handle the truth.
I wonder why /. does not have a section on economics. Isn't it long overdue to have one?
So many stories really belong in economics.
We could discuss what things are worth.
We could point out stories that appear on front pages of various portals and news sites and discuss what really is going on behind the title on them, just like the title I linked to:
Stocks Rise on Renewed Hope for Fed Action
- which sounds as if it is a positive for the economy that stocks rise on 'Hope for Fed Action', when in reality, those who understand can tell you that "Fed Action" means more money printing/borrowing, which implies more inflation and debt, so rising stocks (and rising gold) in this situation means that there is an expectation of yet more inflation, so stocks will go up in nominal terms, but all US holdings will lose more purchasing power.
Isn't /. 'news for nerds' and isn't economy yet another 'nerdy' subject?
You can't handle the truth.
Pinkie in the cheek: "One million dollars!"
Religion is what happens when nature strikes and groupthink goes wrong.
Oh you have to work to keep something up? No shit, never heard of that before. I mean our building we are in, we spend nothing on that. Well except for janitorial staff. Oh and lightbulb replacements. And roof maintenance. And new furniture. And HVAC maintenance. And elevator maintenance....
Seriously, when you buy something the total cost is never just the up front price. You don't plunk down cash and then never again have to spend any time or money for the thing to work perfectly. In some cases it may be direct monetary costs at certain points. The roof at work is a good example, we had to have it largely replaced a couple years ago. Not because of a problem, but because the building is like 40 years old. Had to pay tens of thousands of dollars to do that, which was budgeted. Some of it is indirect, just regular maintenance you have to pay someone for. Our custodial staff is a good example. While some of what they do, like say wash windows, is just aesthetic, much of what they do is necessary upkeep to keep the building in good condition. There isn't a precise dollar figure on each job they do, it is just a general cost that is their salaries. Some stuff has just a time cost, more or less. Like yesterday I dusted off my MIDI keyboard. Needed to be done both because the dust is annoying and because excess dust can work in to the electronics and cause damage.
Why would software be any different? Yes, you have to spend time and money above just the initial cost. You have to patch it, some software has yearly support contract cost (like say RHEL), you have to have support staff to make it work and help people with it, and so on.
I fail to see how this is a "debt" of any sort. It is a "cost" like any other. The more software you use, the more cost ther'll be not just purchase cost but also support cost. This is surprising to nobody who understands how this shit works. This is also why the price tag on software is often not a big deal. Doesn't matter is a package costs $50,000 whereas another costs $50. If the $50,000 ones saves $100,000 in support and other costs (like lost productivity) it is worth its price easily.
To me this sounds like the kind of thing a dumb manager would say: "You mean that price we paid for our software 6 years ago isn't the only thing it cost! Holy shit we have software debt!"
I have five modpoints.
Big deal. I have fifteen (posting Anon so I can mod comments in this story later).
Turns out I don't understand how the system works, after all. I see something in the FAQs about having 10 points instead of 5 for being a good moderator, so I don't know if this is a bug (adding 10 to 5 instead of replacing 5 with 10) or if something changed more recently than the FAQ entry. The last time this happened (the first time I had 15 points), they were still set to expire in 3 days, and the system let me use them all normally. Not sure if I can ever remember having just 10 mod points to begin with.
Anyway, point is, someone needs either to squash a bug or to update the moderation FAQs.
You are BOT. Assimilate, Assimilate!
Wait a minute. I'm a manager, and I've been reading a lot of case studies and watching a lot of webcasts about The Cloud. Based on all of this glorious marketing literature, I, as a manager, have absolutely no reason to doubt the safety of any data put in The Cloud.
The case studies all use words like "secure", "MD5", "RSS feeds" and "encryption" to describe the security of The Cloud. I don't know about you, but that sounds damn secure to me! Some Clouds even use SSL and HTTP. That's rock solid in my book.
And don't forget that you have to use Web Services to access The Cloud. Nothing is more secure than SOA and Web Services, with the exception of perhaps SaaS. But I think that Cloud Services 2.0 will combine the tiers into an MVC-compliant stack that uses SaaS to increase the security and partitioning of the data.
My main concern isn't with the security of The Cloud, but rather with getting my Indian team to learn all about it so we can deploy some first-generation The Cloud applications and Web Services to provide the ultimate platform upon which we can layer our business intelligence and reporting, because there are still a few verticals that we need to leverage before we can move to The Cloud 2.0.
"Its okay, this project/software is using 'internal resources'"
"Say, Jim, would you mind working a few extra hours for the next 14 weekends in a row? I know you're salary, but we'll make it up to you once this project is done..."
And that, my friends, is how you completely ignore hidden costs and justify even the most lingering of projects.
At least at my company, anyway.
I just spent a lot of cash on the Waves Mastering Suite for Pro Tools. In fact any "multimedia" software has a crazy price. I figure they do this to milk media companies for all that they're worth.
If they really want to know what software costs these days they need to be calculating those prices in Rupees, not Dollars
Fascism trolls keeping me up every night. When I starts a preachin', he HITS ME WITH HIS REICH!
The solution is obvious. Ban newlines.
Could impact "software metrics" though. "Three months and you guys have produced only one LOC? You're all fired. We're sending the job to a guy in India who guarantees 20 LOCs per day from each programmer."
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
We run a lot of HP servers.
HP publishes firmware / software updates.
I'm the only person in the department that applies them to any of the servers.
"If it's not broken, why fix it?"
Yeah, it isn't shocking. Mostly because I'm beyond being shocked by anything those idiots do.
If it's physical, they can understand it. They put gas in their cars. They take the cars in to have the oil changed. And so forth.
But software can't be touched. A server running the latest patches looks the same as a server without them.
Unless there is a problem in the software that they are experiencing, they don't understand it.
your cost is the salary divided by the work done.
These kinds of studies often lead to stopping work because it's "too expensive" which leads to your staff sitting doing nothing. It's absurd.
I've seen many small projects which would have 1 to 2 percent improvement canceled because they were not "cost effective" and then the programmers sat there doing nothing for 2 months. You should always let your programmers work on little side projects that they are enthusiastic about as long as they make the big deadlines that you want.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
It seems strange to me that NSF blew 500k on the "cost of software" when the more interesting cost is the cost of data lost to proprietary formats. How many scientific experiments are done, on the NSF's dime, that never get shared with the public in open formats because instrument manufacturers lock in their clients.
as someone above mentioned, the cost of decommission is a real bear, and breaking the lockin instrument manufactures have on the instrument data should be something NSF devotes monies toward.
Gartner's been saying this stuff for 20 years. In practice, the news that software has ongoing costs was used to exterminate any solution except Microsoft from large organizations.
In order to do that, non-IT departments had to be forbidden to install or choose any software at all, or they would have shown an annoying tendency to pick software that worked better for their needs. So IT had to remove all admin privileges and get Purchasing to not permit any purchase orders for software not OK'd by IT.
The larger the organization, the better the argument works. The proportional costs actually drop through advantages of scale; but the absolute cost starts to sound appalling when multiplied by 10,000, particularly when a single, generic, (Garter-quoted) number like $1000/application/year/PC (developed by dividing the whole support budget by the number of apps) is used for all apps, including handy little graphics utilities like XnView.
In theory, a department could prove by business case that some second (third...) solution to the same basic software category within the organization was justified by some benefit. In practice, this was made almost impossible to prove and was limited to graphics jobs being allowed Macs.
It was used to prevent Firefox, for instance; though free, the same "administration costs" were applied to the decision. Pleas that customers could upgrade and manage it themselves by clicking the "upgrade" button were scoffed at; one or two anecdotes of users that got themselves into a mess were multiplied by 1000 PCs to claim that vast costs would be incurred.
In short, this argument was used to take the "Personal" out of corporate PCs. Someone I know discovered there were only a few things he COULD change on his desktop; couldn't even delete the icons for Approved Corporate Software he didn't use. But he could, and did, change one icon to read "Their Computer".
yeah, but was he voted "voted #1 vodka of 2033" ?!
Isn't this the whole TCO thing, all over again?
As a sysadmin, the points about maintenance and downtime really resonate with me. And then there's crap like having a team of 10 to 20 staff (many of them very senior) standing around scratching themselves due to some bug or shortcoming (often just in the UI) that would've taken an hour or two to implement. The cost of *not* writing software can be astronomical. As a more concrete example, I worked in a shop that used LDAP to authenticate a myriad of services (desktop signon, shared volumes, shell access, web applications, mail, etc.), but there were some "glitches" in the LDAP schema and the clients weren't always properly configured to use them anyway. Fixing the issue completely would've taken maybe 30 hours. Writing our own web app to create/modify accounts with a step-by-step set of screens that implemented our business logic for new accounts might've taken 150 hours. But doing something like that would be too costly. Better to eat up an average of 5 hours per week of sysadmins time diagnosing trouble with sign-in to individual services, another 5 hours of the staff's time who were trying to sign in, and the occasional 5-10 hour patch of yak shaving when someone stumbled into the thicket accidentally once a month. And on top of *that* are all the dirty little secrets of employee behaviour to work around the shortcomings of the system.
----
Not to be confused with Col.
Won't that drive up the true cost of maintaining the moderation system software? And adjusting the FAQ could lead to countless man hours used replying to gripes from people that never get extra-extra mod points. /.?
Do you want to drive up the hidden cost of
It's already enormous when you factor in lost work hours...
You have the right to remain sentient. If you give up the right to remain sentient, you will be elected to public office
I think this is an FAQ issue. I've asked a few questions about how things work and made a couple complaints about how far certain information is buried to the powers that be around here. The answer was that the FAQ's are way out of date and there just isn't time to update them.
I have been given 15 mod points 95% of the time for last 4 or 5 months, and almost every day for the last week or so, but haven't asked why after being told how far out of date the FAQ's are. I figure it's a policy change of one type or another. It's either a bonus for being in the top 1/2 of 1% of moderators or something like that, or they are just adding the 10 mod points to the normal 5 whenever you are given mod points now when you're a 1%er.
Costing software by its size ($2.82 per line of code) is not only foolish, it is counter-productive. Much software development effort is applied to making applications faster and more easily hosted on small devices. That is, the effort is in removing code without removing capabilities. How do you cost the removal of lines of code?
Louis B. Mayer once commented about the movie industry, "I realized then that movies are the only thing you can sell and still own." Mr. Mayer lived before software but I'm guessing he'd have been heavily invested in it.
to the coward responsible: present yourself to me, admit what you've done, and i will kill you.
Personally, I take TCO figures with a grain of salt. The methodology is so speculative, coarse-grained and pre-biased that the numbers are, frankly, ridiculous. Somewhat relatedly, a recent Spanish study estimated the total cost of a pack of cigarettes at $150:
http://www.torontosun.com/life/healthandfitness/2010/10/08/15628006.html
... by buying a copy of Applied Software Measurement by Capers Jones. There's enough data in that book to extrapolate the cost.
http://knol.google.com/k/paul-d-fernhout/beyond-a-jobless-recovery#Four_long(2D)term_heterodox_alternatives
Re: "Sadly I think its too late to really do anything about it."
Essentially, rather than create more "artificial scarcity", we need to shift our socio-economic paradigm to deal better with the abundance computer technology can create.
A 21st century issue: the irony of technologies of abundance in the hands of those still thinking in terms of scarcity.
michaelkristopeit 30 is an impostor attempting to steal my identity. signed, therealmichaelkristopeit123456789whodoesntusecapitalletters.
...somebody tell them then, damn.
With most commercial software you end up being on your own in the end, and the end comes sooner than you think. They are greedy after all. I have tried switching to MacOS for some tasks, and now I'm reminded again why I hate proprietary software. One of my latest disappointments there has been GarageSale... a few months after buying it, there's a new version for which I'd have to pay to upgrade. And ebay changes their APIs too much, so I'm afraid pretty soon there will be a change which they will not "support". Being closed-source software it would then become useless. I bought it though because there aren't any good free software alternatives these days, and it saves time when creating auctions compared to using the ebay site... just stuff like drag-and-drop of images, and easy formatting (even though I've got lots of experience writing HTML by hand, it's a boring time-wasting thing to do). Another is Parallels. TWO MONTHS after I bought it, I'm already ineligible for a free upgrade to the next version, in which they supposedly made it much more efficient. So my impression of that company is even worse, but OTOH I do have free alternatives (just not as nicely integrated, I suspect).
Several times I've worked at companies which insist on using expensive software because it has "support", e.g. ClearCase and ClearQuest. Again, they suck in some ways, you can't fix it, and you can get better results with svn or git by just investing a little sweat equity in setting it up and getting use to the workflow. So to me the word "support" is always a weasel word: As soon as someone utters it, watch out... here comes a snow job.
I went a good three-four years without being able to moderate at all, and now I routinely get 15 mod points. Must be allowing me to make up for lost time or something.
If only there was a source for free software which other people patched for you whenever a bug was found.