O'Reilly Publishes a book on the future of software
potatohead
Posts: 10,261
http://www.oreilly.com/programming/free/files/software-paradox.pdf
That's a promo, and I'm not sure how long the link will be good. We have had a ton of open vs closed type discussions here, which is why I decided to link this.
For me personally, the trends in this book are relevant and are going to impact some decisions I make over the next year or two. In a more general sense, our ongoing Linux, Mac, Windows, Open, Closed discussion could use some history and context.
Had to change the title as it implied O'Reilly wrote the piece. The author is Stephen O'Grady. A guy worth paying attention to, which is why O'Reilly has published him. Sorry about that.
Enjoy, and share your thoughts.
That's a promo, and I'm not sure how long the link will be good. We have had a ton of open vs closed type discussions here, which is why I decided to link this.
For me personally, the trends in this book are relevant and are going to impact some decisions I make over the next year or two. In a more general sense, our ongoing Linux, Mac, Windows, Open, Closed discussion could use some history and context.
Had to change the title as it implied O'Reilly wrote the piece. The author is Stephen O'Grady. A guy worth paying attention to, which is why O'Reilly has published him. Sorry about that.
Enjoy, and share your thoughts.
Comments
The money is to be made from businesses and professionals who need more specialized functions. SAP automates your whole enterprise, which is a massive custom project. Much of my work is automating little processes within a plant with embedded controls, which is where a lot of us are in this forum; again we're working for people who know they have to pay for our time to get it done right.
There's no room for another Microsoft to arise because there are no commonly needed functions with large enough markets waiting to be implemented for the first time. Since most people aren't used to paying for software any more many things are now supported by advertising (like search) or paid premiums atop a free base service (like a lot of games). I've seen it estimated that most people would not be willing to pay Google what it makes off of you on directed ads. It's hundreds of dollars a year.
In fact Google may be one of the last companies to find such a new niche. And like all of the existing players, Google and Microsoft seem to be casting about frantically for the next thing that will take off the way their core business did, because the nature of things is that their core business may not be viable forever and then they'll essentially become hollow shells. But being established players with billions of dollars in capital to invest they have a better chance of finding that next thing than I do in my garage -- or, should I do it anyway, of buying me out.
But the real problem is that the number of corners where such opportunities can be found keeps shrinking. Within a few years people won't even have "computers" any more which support general purpose program development. Most people won't need them when you can watch TV and get map directions from free apps on your phone.
Microsoft really needs to find something. Google does not. It's currently in expansion mode on data, and the Nest products speak right to that.
Google defines "search" as "answer" in most contexts. The more data they get, the more kinds and better quality answers they can deliver.
The other thrust of the book is about full stack type efforts. Do what Apple does, what Nest is doing, etc...
Or, look at companies like New Relic, which is all about data, metrics, measure what you want to improve.
Map directions are cool. What backs them is data. And those who have collected it have a huge resource. In that book, two methods are mentioned: collect it, or buy it from somebody who did.
Google maps is much better now. As I'm driving about, it warns me of slow downs, suggests other routes, tells me my ETA, and it does this when I'm in maps, or just as a sort of FYI service now too. All those droid phones reporting location and speed data are accumulating a resource Google is only beginning to exploit.
The future of software is something else. Like the future of mathematics or cosmology. Great ideas do not come from corporations but people. And you never know where they may spring up from. As Bill Gates said in his infamous "letter to hobbyists" the idea is:
"What hobbyist can put 3-man years into programming, finding all bugs, documenting his productand distribute for free?"
That is to say, "we will program and you don't have to worry your little heads about it".
True enough at the time perhaps, although I doubt it. Today we have a huge human population many of them connected via the internet. Some of them, a huge number, driven to create software. Which they do. At the same time we have huge money making corporations whose primary business is not creating software, think Apple, Facebook, Google, Linkedin, Amazon. But they collaborate to get the software they need created and perfected. Think Linux, Clang, GCC, many other things.
At the other same time people have come to realize that old fashioned word processors and spread sheets and WYSIWYG desktop programs oriented toward printing on your laser printer are irrelevant in this web connected world. Bottom line: Software has no value. Never has had. It's like asking money from people to calculatethe area of circles. Might work until they figure out that a = ? * r * r for themselves!
Full text of Bill Gates odious letter below:
This: Software has no value. Never has had. It's like asking money from people to calculatethe area of circles. Might work until they figure out that a = ? * r * r for themselves!
Is increasingly right on, save for some niches where software actually does have value. Mechanical CAD is one such niche. There are others, and they center on code bodies that take very seriously large amounts of people hours to create. And to follow your example, computing the product of geometry and topology operations works, until people can figure them out for themselves. Some of those problems can take most of a lifetime to learn to understand, and they can take teams very large amounts of time to resolve. A software solution to all of that can and currently does have very significant value, and...
...is the growing exception.
If it (software) has no value, why do software then?
And that discussion is what lies in the text. Did you actually read it?
My post above was not inclusive enough to spark the right discussion. I apologize for that. This one does frame up what the book is about.
And this is from O'Reilly. Good people Heater. Sharp. Open advocates and friends.
Sure I understood the text. Of course I read it. I have lived through most of what it describes.
Let's look at it another way:
The computers of the early days were huge, power hungry, and expensive. There was no value, comparatively, in software. People would invest huge amounts of money in such machines if they thought they needed one. Think the military or early corporate business use. Those people already had the talent on hand to make use of the calculating power those machines made possible.
By the time we get to the first micro-computers the cost of the hardware was almost zero, comparatively.Normal people could by a computer to take home. Only one problem, their machines were useless door stops unless they knew what to do with them. An opportunity that Bill Gates and friends astutely observed as a money making opportunity. Software to make computer owners door stops useful. Binary only distribution no knowledge sharing mind you. So far we can compare the developments in computers to other industries of the past: The inventor of the steam engine, James Watt, could not work as an instrument maker in Glasgow. There was secret knowledge that those guys were profiting from and vigorously keeping to themselves with their "Guilds"
Before that the best glass makers in the world lived on an island off Italy. Those who worked there were not allowed to leave the island for fear that the secret techniques leaking out. Oh, and what about the masons that travelled Europe contracting their services to anyone wanting to build a Cathedral? But what now? The hardware is essentially free. But we also have billions or people on this planet many of who are smartand like to program. The hardware cost tends toward zero. The software cost tends toward zero. Heck, even I worked for some years for Racal on a PCB CAD package that sold for then thousand dollars a seat and more. Today much the same can be had in the open source KiCad. You are right. There are still major seams of income to be made in software. The CAD systems you mention a good example.
I know a guy that works for company that provides ship simulation software. Did you know that every ship built in the world has to be simulated with such software before they can get insurance and go to sea? Turns out my friends company is the only company in the world that provides "approved" simulation software for such insurance!
What I claim is that software has no value. What has value is the "secret" knowledge encoded in that software. As more and more of that get's out the software price gets lower and lower.
You ask "If it (software) has no value, why do software then?" Well the Apples, Facebooks, Googles, Amazons, and others of this world are making such huge amounts of money that hiring a few guys to create the software that makes their business run is essentially free. The software has no value, and for that reason these companies hire people to work full time on free and open source code. They collaborate on this software infrastructure with what you might think are their competitors.
Why? Because software has no value
You're confusing things.
All these companies employ lots of people as coders. Apple wouldn't be where it is unless it developed it's OS starting with Macintosh through OSX. The same goes for the other companies.
Yeah I understand you're really into the communal thing of working for free. I just wonder how you sell a education in CS or EE and then telling the students after they graduate they have to labor for free.
Really if SW development is going to the work for free model, I'd say young people are better off not studying to be a coder or engineer since they will be treated like chumps and patsies and end up in the soup line.
No. I'm not confusing things. What you say is true regarding Apple an it's software. They keep their secret sauce secret.
But don't forget Apple's operating systems are based on BSD/Mach kernel, whatever. They are developing the Clang C++ compiler and surrounding tools, in collaboration with Google and others. Their web browser is based on webkit which started as the opensource KHTML developed for KDE. The same webkit as used in Chrome. They have recently open sourced objective C.
Given the huge revenue and profit margins of Apple the software is essentially free. No value.
Please don't mistake what I say as the "communal thing of working for free". I am not imagining hoards of basement dwellers hacking on code for Linux, GCC and so on for free. That is not what is happening here.
That "community" is a community of the likes of Apple, Google, Amazon, IBM, Samsung and a host of other for profit companies who happen to realize it's in their own enlightened self interest to contribute to the pool of infrastructure software. They all benefit from it. Else they would not do it, right?
I never said anything about people working for free. Who would do that? On the contrary I believe these people are working for said Apples and Googles and many other businesses and earning good salaries.
And finally, if a young person is only studying a science or mathematics or engineering or computer science only because they hope there is a good money making career in it I would advice them to quit. Clearly their heart and passion is not in it and they should not be there.
Oh, and finally finally, why should a typical software engineer expect to be getting a bigger salary than a book keeper or accountant or many others?
That means the cost of producing the actual product they sold, software, was maybe 10% of their running cost.
Seems software has no value if you are the guys creating it
It's not the problem, but the implementation of a solution to a problem that generates the value. And the problem has to be of a scale to enable a return for the implementer of the solution. That return may be in money or ego gratification (FOSS?, OBEX?); watch the scramble to be the first to find/fix a bug. Creds and reputation, which can be used to leverage into a paying opportunity later.
There are many overlapping solutions for nearly every problem under the sun. The skill of the implementer in creating a solution of value to an end user makes the value. Otherwise all software solving a problem would be priced the same as it could be said to solve the same problem. But we know better. Which is why in my own opinion that some people/companies will continue to make great money and others will be placeholders in line at DES/Workforce commission offices.
FF
Heater, as to why one job should get paid more than any other in a free society, it's all about perception of value and what someone is willing to pay. The classic American example is $42K/year for elementary school teacher v.s. $43Milion for an N year contract for a well known sports star. That is why. No logic, just what are people willing to pay for. Think Gauguin or common street artist.
Heater, as to why one job should get paid more than any other in a free society,
I did not ask that question. I'm fully aware of why salaries and incomes vary so much. I'm not proposing a command economy that dictates we all have the same income.
That's not quite what I was getting at, when I said "why should a typical software engineer expect to be getting a bigger salary" I was referring to Rod's statement "I'd say young people are better off not studying to be a coder or engineer since they will be treated like chumps and patsies and end up in the soup line"
What I'm seeing now a days is a lot of young people signing up to study CS or Eng or whatever for years solely in the expectation that it is the road to a nice job and big salary. Many have no passion for the subject and perhaps will never be much good at it. Now a days they have to borrow huge amounts of money to do something they don't like much. This seems like a miserable situation with no guarantee of the outcome expected. (Contrast to times past when kids would quit school and get a job, there receiving on the job training, apprenticeship, they were paid to learn).
So I was only suggesting that perhaps Rod is correct, they should not do that unless they are passionate and "driven" about the subject and don't much care what the outcome is.
Encoding knowledge into software is going to prove very useful. The very first program I wrote and sold did that. It was for sheet metal layout, and I took the old school techniques, applied some math to bring them up to modern utility and made a bit of software that helped people get flat sheets from their 3D models.
The math I did streamlined and to some degree improved on what is common knowledge. Or, maybe not so common given how often I teach layout, but it's not hard to find in any case. What is hard is actually understanding how to apply that knowledge and software helps big here. Put an interface on something, make it data and user friendly and there is a nice time and often process savings possible.
Interestingly, that kind of thing could be made open quite easily. Over the years I never saw that happen. The audience for it is just a bit too small and not generally populated by people with software skills. Those means and methods got sucked up into modern Mechanical CAD systems, which by the way, do get it wrong on some specific use cases. (mine does not, and that's because I bent a lot of metal in the early 90s')
One of the reasons I put that text here was to highlight the big battles we all enjoy and are forced to be a part of. I see the FaceBook, Google, and friends in a way similar to how I see Disney. Disney took the works of the Brothers Grimm, which were public domain due to expired copyright terms in effect back then, and they produced great color TV programming and movies from them, kicking off the empire we know today.
Disney doesn't want to give back though. Not ever. We all know that story. Extend copyright just long enough to keep Mickey off the public domain register. (and that's theft from the public commons a lot more than infringing on a copyright is theft to the rights holder... discussion for another day)
Today, those Internet empires build off the open tools, and they provide services, and sometimes they give back, sometimes they don't, and sometimes they are friendly with the data, sometimes they aren't, and one can never quite be sure what to count on. Despite this, the body of open software continues to grow and the concept of use value exceeding the cost of learning open tools continues to be a favorable equation too.
All of that bodes well for people like us! We want to make stuff, or make stuff work, or solve a problem, improve a process, etc... And that will consist of taking knowledge, maybe secret, maybe general, maybe specialized, encoding it, and then putting an interface on it, making it data friendly, and process / machine friendly too. Ideally, plug and play, but often just enough of that to make installing, configuring, provisioning, etc... possible and practical.
Propeller 2 is relevant here too. Whatever else it ends up doing, it's going to excel at moving data around, streaming it through software, delivering interfaces, etc... And it's likely to do that without a lot of dependencies too.
I find that extremely interesting.
No OS, for example. That gets rid of a lot of stuff, and it does wall off quite a bit of stuff too. But, the good part, being able to encode knowledge, provide interfaces, stream data, will be solid and easy to glob together in ways that people find useful and productive.
Back to the more general commentary...
It's a good gig and getting better. That's the short message I found relevant. And the idea of open as a general purpose tool box only makes sense when a lot of people can use the tools too! IMHO, this is great for people like Parallax, who happen to excel at and feel passion for enabling ordinary people to do this stuff.
Go to school and see 100K debt?
Not on your life! I'm largely of the school of hard knocks. Great school. It's cheap, brutal, effective. All you gotta do is pay attention when the bell rings and you are very likely good to go.
(the idea of higher education being associated with such debt is politics, and maybe worth a talk somewhere else over a beer or few, but in the US, that's reality right now, so I put it here)
Software can be learned by anyone willing to write it. And with such a huge market cap on the horizon, learning to do software and in particular, the ability to couple it with real world things, is rapidly becoming a baseline skill in the modern world.
Few will author computer code and sell that to be rich. Maybe in games, or some new Internet thing. Sure.
But a whole ton of people will author computer code to get stuff done and live nicely and enjoy what they are doing. All good.
Regarding software value: I think it's more important to understand software has value in context, or in tandem with something else. At the core, it's an enabler. And the book does go through this cycle, highlighting back in the day when software was pretty much free or nominally charged for to enable the sales and use of hardware of some kind, through the time we are leaving now, where software itself is sold for nearly all purposes, and the thrust of the text is about what the next turn of the crank will or could look like.
I have both a technical and a business outlook on things and I find some understanding of macro level trends coupled with specific domain knowledge tends to be very useful and profitable. I put this here for the macro level analysis it contained. Not so much the ideological, but the practical. As I've written here many, many times, having tech skills is great!
Being able to talk to others who do not is seriously enabling, and this text is offered with that idea in mind too.
Heater, I think I see what you were getting at. Happens all the time in medicine, especially with legacies. I really wish they would have opened the doors to the med schools a lot wider when we did the health care reforms. Flood them would have been my preference. To many good students with the level of passion to exceed never get the chance, and connected legacies who look for the buck get the preferential treatment. So not as much competition, more mediocre practitioners. And they won't really make the great paychecks in the end. To be great takes the passion and drive to be the absolute best and the willingness to actually work towards that end. I guess that is where you were going with your question.
Not graduating is the point! Embrace it with both hands, do stuff, fail, do more stuff, succeed, wash, rinse, repeat.
Both your and heaters comments on overall financial expectations and value to society are spot on BTW.
Basically it has conned a generation into selling themselves into servitude. Desperately working to pay back their debts. The corporations, who benefit from this, don't have to do anything or invest anything. The governments have washed their hands of the problem. Our academic institutions have become businesses.
It's the same idea as the human traffickers bringing people to the west and then having them work forever paying back what the trafficker charges for the service.
Whatever happened to the idea that we as a society would probably benefit by finding the young, bright, intelligent, passionate, enthusiastic kids and making it possible for them to follow their passion. I.e. pay them. No matter where they come from or what their family wealth and background is?
Lots of ongoing self training and the occasional employer paid training. I have been know to take a PTO day to go to some seminars that Avnet puts on like the BLE Pioneer kit presentation; the reason I posted about the Cypress PSoC and PRoC devices. They have been free so far and very well worth the day to learn a new tool/trick. Creds @ UHK
Think about this... Spreadsheet software...
It takes a huge number of man hours to write and debug the complex spreadsheets of today. Now MS made squillions out of Excel (and Office). But they then were challenged by Lotus123. MS answered by lowering the price till they squeezed them out of the market. Then MS raised the price to gouge again.
Now it needs to be said its not likely that a hobbyist is going to write a new spreadsheet and open source it, yet there is no market for it otherwise.
But what happened is that the following a number of acquisitions, one of the new owners of the Lotus suite open sourced it. Ultimately this caused Office price to again be reduced.
So IMHO the only reason we have Open Office is because of MS tactics, not because of hobbyists.
CAD open source has come about because this software is very expensive. So some cheapie and far less feature software was released. When some of this became open sourced, many hobbyists jumped on and cord various parts until the open sourced versions had many usable features. The expensive software still exists for specialised purposes.
It's interesting to see the various things that come into play over time.
But the hardest thing is to predict how the future will play out.
So IMHO the only reason we have Open Office is because of MS tactics, not because of hobbyists
Yes and no.
I'm not sure Open Office ever had anything to do with hobbyists. Open Office originated as StarWriter by Marco Börries. A commercial product for the old 8 bitters and CP/M. Later adapted to 8086 for Amstrad PC's and so on. Growing up into a office suite for IBM's OS warp. Licences for StarWriter/StarOffice were 50 dollars or so.
Sun bought the company StarDivision back in 1999 for tens of millions so as to get an office suite for their machines. It made sense for them to give it away for free so as to encourage sales of their machines.
As we say, software has no value
Later Oracle bought Sun. StarOffice became OpenOffice and here we are.
So yes, this is all a cut throat commercial battle to get people off of MS and onto your machines. There are not many hobbyists in this story. I wonder how many such "hobbyists" contribute to OpenOffice/Libre Office now a days.
By the way, has anyone ever managed to use OpenOffice to actually create long document with any degree of structure and formatting? I find it almost impossible.
When I need a spreadsheet I use Gnumeric, because it's easy, fast, and creates small files. And it can export to LaTeX tables. And I can actually use it without training, unlike Excel. Gnumeric is Free software, yes open source, and in this case a better product as well, to me.
Hey, I remember playing around with that Borland spreadsheet source back in the day. Great fun.
As far as I can tell spread sheets are great for simple calculations. Use like a souped up calculator. When things get big and complicated it's better to start thinking about doing it with a proper programming language else you are going to get into an unmaintainable mess of bugs.
Of course the extendability isn't all good - the tables people send me are wider than the screen by a factor of two or worse.
By the way, has anyone ever managed to use OpenOffice to actually create long document with any degree of structure and formatting? I find it almost impossible.
Well, I (and others) can never figure out how Word works with numbered paragraphs! Formatting is a pain too!
I should take the time to figure it out, but whenever I have had to do it I am in a hurry. I absolutely isn't intuitive!!!
SuperCalc and others were quite simple. But Excel has become so powerful as well as bloated. My wife and her boss use massive spreadsheets. My wife hasn't programmed for many years, but her boss programs almost daily. Excel can things simply that would take years to program because of the ease of modifying.
Often power users import data from various places and use excel to reformat and calculate and produce complex results. Some spreadsheets are reused regularly, and some only once or twice. And excel can be used by non-programmers or self taught business people alike.
As for screen size, my wife and her boss both use a pair of 27" LCD Monitors each.
I have written VB code for Excell for a commercial software package.
In many ways, it's like CAD software. How do you charge different prices to small users versus power users while keeping the one package? In hardware, this was done easily back in the day when computer companies maintained them. Rip out a capacitor, a board, or apply a software patch and you computer, printer, hard drive goes twice as fast or has twice the storage, all for a handsome price of many thousands of dollars.
So the students, who lack the basic Reading Writing and Arithmetic skills, are "given" a diploma upon graduation, enter the workforce, fail on the technical stuff, and are demoted to being a "people person". These people end up being our bosses, making decisions about the companies we work for - civic & government leaders, making economic decisions we pay for, environmental mistakes we live with, defense & security we rely on, laws & policies we have to deal with.
Sadly it looks like that. They just need to step on the market place and be good at jelling 'I am the best'.
Mike