When a good idea fails the loss is not just that idea, the failure scares away potential investors from anything resembling that idea; consequently,
innovation suffers and everyone pays the price. The software industry is especially good at killing good ideas, and Usman Latif’s article “Why Good Ideas Fail” discusses the reasons behind this terrible record.
Editorial: Why Good Ideas Fail
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I agree 100 percent. This is relevant not just for apps and OSes either, but gaming as well. So much stuff sucks.
Remember how diverse computing was before Microsoft?
Actually, I meant how diverse computing was before “Windows 3.0”
Anyways, MS has their fair share of good ideas, but damnit they need to go away.
A note to our viewers: The following post contains a Lisp rant. Continue reading at your own discretion
He hits many points right on the head. It seems that the smallest things are enough to kill products. The fact that companies are conservative in their investments and wary of potential failure does not help. My favorite example:
In 1993-94, Apple’s Cambridge development lab invented Dylan. They basically got together a bunch of expert language designers to make a new language for their Newton product. When the first previews were released, Dylan was impressive. It had an incredibly sophisticated IDE that displayed Apple’s GUI prowess. It let you edit code while you were working on it. Code wasn’t split into files. Instead, you got a graphical view into all parts of your program. The language itself had most of the power of Lisp, much of the simplicity of Scheme, was fully object-oriented like Smalltalk, and generated code nearly as efficient as C. It was dynamically typed, but supported type declarations for increased safety and speed. It supported multiple syntaxes, to disconnect the semantics of the language from the text of the code.
By 1996, Apple cancelled development on the project. Why? For many of the same reasons listed in the article. Apple ran into financial trouble at about that time, and dissolved the Cambridge lab. They also marketed it terribly. Apple pissed off the Lisp users by abandoning the option of a parenthesized prefix syntax and pissed off everyone else by designing a Lisp-y language.
Now the interesting thing is that these ideas are rediscovered over time. Lisp, for example, pioneered garbage collection in the 1950’s, and Java made it popular in the 1990’s. When C# 2.0 comes out, it’ll finally get lambdas and closures. Python is sneaking dynamic typing into the mainstream in the guise of a scripting language. Apple, more forward looking than most of the behemoths, has gotten a whole market of developers to sample the dynamically-typed goodness of the Smalltalk-inspired Objective C. Microsoft is discovering the power of allowing domain-specific languages in Xen. Maybe sometime in 2010, C# will get something resembling Lisp macros
Open source also offers a way out of this viscious cycle. Dylan is happily plodding along now as an OSS project. Unlike commercial companies, open source projects are not constrained by having to satisfy the bean-counters, and can thus persue a good idea for as long as they desire.
That should read: “let you edit code while you were running the program” not “let you edit code while you were working on it.”
This is why Linux will prevail. Linux has time on it’s side. Linux achieved most of it’s progress while having very little corporate support and now it will accelerate even faster with the support it has. Even if all corporate support dry’s up Linux will continue marching own because it was built by people for people to use and spread. Will the corparate support helps and is welcomed Linux will not go away if that support goes away since it was never dependent on it. Will comercial Linux’s die off ? Yes but they can easliy be replaced. It’s the huge grass-roots development and testing foundation that Linux has built which will keep it around for as long as people are interested in computers and OS’es and software.
Software users tend to stick with software they are familiar with. This is not irrational behavior; software is complex to learn and use productively, and people prefer not having to relearn computer skills over and over again. Most people are inquisitive about new products but can not see straight away if these products carry any productivity benefits. Moreover, the utility of new functionality often becomes apparent only after individuals get accustomed to it.
So true, it’s why it’s hard for many companies (even include families) to switch the desktop. It’s not that simple to switch Windows to something like Linux (KDE/GNOME) or MacOS X; it requires a lot of time to get people learn, understand and to get use with it.
Create the clone is kind of the only way to go and move slowly change (de-clone) interfaces, features, options and etc at the each version to get people slow notice the changes and easily to get use.
Excellent point. That may just be what Microsoft is doing with C#. Make something familiar (Java clone), then add to it over time.
Familiar shmamiliar, users didn’t switch to BeOS cos it didn’t have the apps they needed, developers didn’t switch because there weren’t enough users. A kind of catch 22 but one that could have been overcome given the right strategy to get developers on board.
“So true, it’s why it’s hard for many companies (even include families) to switch the desktop. It’s not that simple to switch Windows to something like Linux (KDE/GNOME) or MacOS X; it requires a lot of time to get people learn, understand and to get use with it.”
Companies will never change simply because a product is superior. The change has to entail benefits (profits in the business world). And these benefits have to outweigh the risks in the long and short terms. There are many examples in history that show superior products losing to inferior ones, at least technically. VHS vs beta for example. Beta was better in every respect than VHS was except for two things. 1. Sony royalties and 2. Shorter playing times. Porn industry support also helped VHS. Familiarity and learning time have nothing to do with it. It’s all about the money.
Also, it’s always the programs that support the os, not the os itself that urges a company to move. When you move from windows to linux for example, you’re also choosing to move away from windows programs. So progs for linux have to be superior enough that the competition won’t ever catch up and be justified through profits. Windows has the largest app support of all, and it’s easy to find somebody that has experience coding for windows if you’re needs are specific.
“Software users tend to stick with software they are familiar with. This is not irrational behavior; software is complex to learn and use productively, and people prefer not having to relearn computer skills over and over again. Most people are inquisitive about new products but can not see straight away if these products carry any productivity benefits. Moreover, the utility of new functionality often becomes apparent only after individuals get accustomed to it.”
This may well explain why OpenOffice.org has not has much success so far. I have used it, and so many other people have adopted it as thier office suite of choice, but it is not a very strong contender to MS Office. Hopefully, it will gain popularity over time, forcing MS to actually compete for office suite share once again. I think that MSO has reaced a plateau, they have added so many features that they have to improve dramatically to convice users to upgrade. Since MSO 2000 does everything that is nessesary, I feel no need to update to MSO 2003 even though I beta tested it. Granted the Outlook spam filter would be handy, but I have added in SpamBayes, so that is not even relevant. MS will have to add a lot more to convice me to upgrade to a noher one ot thier Office products.
The only reason I am not using OOo is because there is no intigrated mail client Even if it was a simple one, it would be nice to have intigration with the rest of the office suite. Wy should I have to use OOo and Thunder bird, when I can use MSO 2000, and have all of the individual applications have the same UI structure? Makes little or no sence to me. It would be better if OOo would intigrate a mail client that is comparible to Outlook, snd then I would convert over to it full time. I can live with the other annoyances/differences, but I cannot stand having a non-intigrated office suite.
Yeah because Linux has never ever killed off diversity.
Ideas for it to be really successful carry a feature that allows management to report an incremental trend based on the 90-day milestone. Failure to comply with this method is simply burning investor’s money.
Linux is diversty….if you don’t like it then fork it !
Very nice flash tour from Quantrix. Also nice that flash now works well on Linux.
I’d never heard of Quantrix. It is a tough environment to introduce innovation in. The program is not a plugin to Excel but an alternative written from scratch. E.g. a lot of work and it’s ‘against’ Excel rather than with Excel. And if it should catch on then MS will adopt the idea.
The odds are not good for innovation.
Because MS got too big. MS should have been split up – just like the baby bells. Maybe even that is not enough. To allow others and their innovations a place at the trough. But the competition would lower prices and therefore tax revenue and in these stagnation times government is not going to go there. On the other hand they are against Oracle engulfing PeopleSoft even though Oracle still wouldn’t be as much a monopoly as MS is. But… this is getting off topic.
EA, SplashDamage, and developers of Call of Duty should team up together and make cool ww2 game with vehicles cool long maps, hi res models, and mind blowing objectives
I can’t wait to have Quantrix on BeOS
Companies will never change simply because a product is superior.
Yes that include, so good point and I agree.
Familiarity and learning time have nothing to do with it. It’s all about the money.
I disagree. The familiarity and learning time have to do with it. I have been there by teach my grandparents and other people; it’s not that simple to get them switch or make the fast change. There are more home users than companies, so it’s money of OS/apps. In the business, the learning time required the money; time is money. 🙂
The author of this editorial has some good points but he misses the many examples of innovation in the commercial and open-source/free-software communities.
Apple’s OSX has many great new features inclusing Expose`, iTunes, SVG UI …
Java is in innovative and pretty successful language that brought virtual machine languages, garbage collection, and cleaner object-orietation to a mass audience.
Mozilla pioneered pop-up blocking, tabbed browsing, …
Gnutella pioneered P2P file sharing.
Debian’s APT package system is a great example of open-source innovation.
These projects range from a weekend’s worth of work for a single developer to huge products that took millions to bring to market and they all are successful or influencial examples of software innovation.
The pessimistic tone of the editorial is unwarranted. Not every good idea will succeed but a heck of a lot of them will. Don’t worry about software innovation it is alive and well.
iTunes is hardly innovative. Its an MP3 player for god’s sake. And OS X doesn’t have an SVG UI, and its Display PDF UI isn’t innovative either. Sun’s NEWS had a Postscript-based UI, and of course, NEWS was a commercial failure.
Java is not at *all* innovative. It might have brought these features to the masses, but garbage collection and virtual machines are as old as dirt.
Popup-blocking and tabs for browsers are such minor innovations I don’t think you can even count that. Popup blocking is kind of obvious — popups are annoying, so lets add an option to disable them. Tabs are just a form of MDI, and have existed nearly forever. It didn’t take a stroke of genius to apply them to browsers.
P2P was an innovation, but it was poineered by IRC more than anything else. And who still uses IRC?
APT is pretty innovative, but its just a package manager. Perhaps what’s more innovative than APT is the whole idea of having a huge, well-oiled machine packaging a central repository of software.
The guy is right about Linux. It has evolved by itself, it doesn’t need the backing of the Big Players in order to exist. I know that the Big Players contributed to it in many ways with filesystems and so, but even without that Linux would still evolve. Of course that having big names as allies favoures the acceleration of this growing and maturing process. I think that in almost 15 years Linux has achieved quite a lot. Imagine what would have been if Linux would have had this kind of backing it has now from the begining.
One of my biggest regrets is that BeOS failed to make it. It showed so much potential.
Anyway, it’s not like Linux is the ideal OS, in fact many will point out so many bad ideas in the internal design that you would call it anything but ideal. At least it is design with more security in mind compared to “another O$” we all know. But we need competition in the consumer OS market. We have good examples of how competition helps innovation and squeezes the talent out of the HW/SW engineers: Intel vs AMD, Nvidia vs ATI, etc.
My 0,02 eurocents
Waiting impatiently for the second part!
Stray: Remember how diverse computing was before Microsoft?
Stray: Actually, I meant how diverse computing was before “Windows 3.0”
I do. It was amazing. I almost pity the younger generation, but then there’s all this current massive amount of free software which gives an approximate idea. We really should do some kind of “program paleontology” to review OSes and tools of the past. And games, too, why not? 🙂
Usman: Linux exists because Linus wisely chose not to be ambitious.
That phrase alone made it worth reading the entire article. I think this still makes Linus an admirable guy.
That may just be what Microsoft is doing with C#. Make something familiar (Java clone), then add to it over time.
Microsoft was always very careful to clone itself which is one of their best strategies. Even WinXP can be easily changed to look’n’feel similar to 95, 98, and 2k. I think it was even possible to let Win95 look like Win3.x
On the other hand, one can always criticize such behaviour as adding bloat and lacking innovation. Obviously, there’s a trade-off between these possibilities.
I’m not sure if this the same spreadsheet I remember hearing about. Maybe it was another one. The problem was that it wasn’t good for what ordinary people used spread sheets for. It was good for people who thought a spreadsheet was a simple database, but most people only used a spreadsheet for displaying stuff in simple tables. The problem was that it was ill suited to that kind of thing, and was largely rejected. I don’t know why they didn’t evolve it by adding back those most basic features.
I think that an issue that was missed is that often the developers don’t really realise what they are aiming for. Even now with big business software the end product is a rather hazy concept that changes from month to month or person to person.
Often projects lack direction. The thinking that got people started is often different than what they end up with. In many ways software is a tiny part inspiration, and a lot of evolution. The WWW is a big example of this. It is unlikely that anyone really envisaged the final result when they first started fiddling with hypertext. And its not like it arrived overnight.
One thing that seems to define OSS is a lack of where it is going. Many projects are just reimplementing the wheel(office software). Other products seem completed and are only really tweaking and implementing the current plan, not innovating or planning a decade ahead (Apache, SMTP servers, XFree).
MS has the same problems. Most of its software development is evolutionary rather than innovative (despite what Bill says).
I’m not sure what to say about Apple here. I would hazard to guess that Apple has a great design plan (either that or they are very lucky), but no marketting plan to speak of.
I’d say that its the exception rather than rule to produce software that fulfills an immediate niche, often its a matter or finding a market and tailoring yourself towards it. I agree though that it will nearly always take 10 years or more to break into an existing market in a noticable way.
And for the twit listing software above
Opera really started tabbed/MDI browsing.
Napster really started P2P sharing. (How could you ever get that wrong)
let’s not forget the saying “the road to hell is lined with good intentions”
Agreed!
I got into MS-DOS compatible computing in late 1984 or early 1985 (sorry, getting old, can’t remember any more). Back then, there was a plethora of software for the user.
I watched, over the years, as one company after another was taken down by Microsoft. Some may say this is the way of capitalism, but at some point I believe it grew beyond that.
As large as Microsoft is these days, they need to be regulated. Heck, the government regulates every other monopoly.
I’m all in favor of Microsoft continuing to produce their product. My only gripe with them is that they kill good software companies in their struggle to capture 100% of the market.
Wow. Quantrix looks amazing. Not something I would use too often, I’m sure it would be great for businesses. This is a similar idea to using attributes instead of directories for file organisation, or relational databases. That is, you enforce no exact structure so you are free to invent new structures and views without a ton of manual work, and to gain insight into data.
Stray: Actually, I meant how diverse computing was before “Windows 3.0”
I do. It was amazing. I almost pity the younger generation, but then there’s all this current massive amount of free software which gives an approximate idea. We really should do some kind of “program paleontology” to review OSes and tools of the past. And games, too, why not? 🙂
I got in to computers around the Win 95 era, so I can’t say I ever saw that diversity. Personally I often read about computing history, and the narrow view of software that my peers hold worries me a lot. They think I’m crazy when I get nostalgic about Acorns we had at school.
Go to http://www.old-computers.com, browse by year starting at 1995, going backward in time. This may be just an anomaly specific to this site, but notice how there are gradually more computers the further back you go? The PC killed them all.
I watched, over the years, as one company after another was taken down by Microsoft. Some may say this is the way of capitalism, but at some point I believe it grew beyond that.
Read the book “Barbarians led by Bill Gates” to find out what
Microsoft has done in the past to outflank a competitor or suppress a technology that they didn’t control.
I had never heard of “Barbarians led by Bill Gates”. It sounds interesting, and what the heck, we happen to be heading to Borders today, so I will check it out. Thanks!
Re: @ Luke McCarthy: By the time Windows 95 came on the scene, Microsoft had already established itself as the dominant player in the personal computer market.
I’m also glad I got into Linux and the BSD’s. To me, they brought back a lot of the “fun” of computing I experienced in the early to late 1980’s. Before Microsoft became the defacto standard.
Yep, the PC killed-off all the other computers out there. Those of us that are into retro computing (I have a nice collection of Tandy and TRS-80 computers) can still relive the “early days”. 🙂
“This may be just an anomaly specific to this site, but notice how there are gradually more computers the further back you go? The PC killed them all.”
If you look at the number of car manufacturers in the 1920s you’d probably be surprised at the variety of cars and ideas. Consolodation happened in that industry because cars, like PCs are fairly expensive to design and build. Economies of scale can be taken better advantage of by people who can afford to build big assembly lines. I think consolidation is probably a natural side effect of captitalism in new industries.
This book is awesome…. I’m loving every paragraph. I only wonder… did this book accomplish anything? Ten years later, everyone still thinks that the best system to model new systems around is Unix. Hello Apple??
“Mozilla pioneered pop-up blocking, tabbed browsing, …”
Opera, a commercial browser, had tabbed browsing long before Mozilla.
“Gnutella pioneered P2P file sharing.”
No. It didn’t.
The only OS I seriously use at this point is linux. I primarily use free software, and love it. Don’t attribute the “innovations” of others to it.
I can point to many, many cool things which are linux or free software related; without a lot of research, I can’t say which are innovative.
Figures. The Unix Hater’s Handbook is hosted on a Microsoft site.
The competition in the hardware world was killed off by a mono-x86 software world together with cheap x86 hardware. These computers were just so much cheaper. Wintel got popular at a right moment and at a cheap price. As we’re all aware here, that doesn’t say a dime about either of it’s quality.
The author claims FLOSS can’t innovate and that FLOSS developers can’t be blamed. There are for certain innovations in FLOSS; in order to clone an existing, proprietary application, one has to make features which already exist in that proprietary application. And the UI for example needs to be user-friendly. Well, newsflash: proprietary UI makers like Microsoft (Windows) and Apple (MacOSX) already are The King regarding that; they have to, they aimed at being user-friendly, that attracts new people which existed in that time. So why reinvent the wheel? Else, FLOSS devvers wouldn’t have any user-base at all (which is important)(and which some applications relatively don’t plus they’re not very known). That makes it harder to identify an innovation, but i also partly agree: it’s how FLOSS has been developed the past 20 years, it has always been like a “race against your own shadow”. However when the quality is (near) equal, or succeeds to proprietary software it’s time for those vendors to start peeing in their pants. Yet another reason i can think of is that research costs are expensive.
Anyway, it can’t be by definition. Think about the following: when some company would release their innovative, proprietary software -which happens sometimes- then it is “a FLOSS innovation”.
You can name Napster, you can name IRC, but none of these were decentralized like Gnutella was.
FLOSS it’s most unique feature is that it resembles the zero-distribution cost (or zero marginal cost). Which is basically what it is about imo. But another unique feature is that it is multi-platform, and in many cases if not, can be ported. That goes futher than Linux, or Linux distributions. You can already see proprietary OSes benefit as well (GPL can’t protect everything), including those which run not on non-x86 platforms.
Imagine you are a newcomer in the hardware industry. Wouldn’t you love all these free (both apply here) goodies? It forms an awesome base of applications which already run on your platform. You don’t have to code it, you only need to port it! I think that’s one of the reasons why a hardware firm like SGI likes FLOSS. It doesn’t hurt their hardware business plan, it enables the industry more diversity; indirectly.
Amazing article. my only hope is that, as has already been mentioned, the open-source movement, which doesn’t have the emphasis for an idea to be profitable immediately as private interests do, will be able to bring more innovation to the main stream where private interests have failed. Given, there will be less press, but word of mouth can be powerful.
This book is awesome…. I’m loving every paragraph. I only wonder… did this book accomplish anything? Ten years later, everyone still thinks that the best system to model new systems around is Unix. Hello Apple??
Most of the stuff in this book are grudges against particular applications. There is precious little wrong with the Unix “philosophy”, if you will. There is plenty wrong with the Microsoft (and Mac!) philosophy.
I too think a lot of the stuff in the UNIX Hater’s Handbook is petty (and obsolete), but there is a kernel of truth in there. Contrary to what people may believe, the UNIX Hater’s Handbook isn’t a Windows or Mac lovers’ handbook. The alternative to UNIX that these people were espousing were systems like the Lisp Machine and the TOPS20 system. The Lisp machines, for example, were extremely advanced. They had hardware assistance for things like dynamic typing and garbage collection. They didn’t have a performance-sapping userspace/kernelspace boundry — instead, everything ran in one Lisp image, and saftey was provided by the language rather than the OS. Courtesy of the Lisp “read” mechanism, LispM programs did I/O in terms of objects rather than raw bytes.
In retrospect, many of the advantages of the Lisp Machines (except the userspace/kernel arrangement) can be provided on top of a simple UNIX foundation. Indeed, UNIX itself makes a nice substrate for more advanced userspaces. However they point they were making, which is still somewhat valid to this day, is that UNIX->user interface is too low level. UNIX is low-level by design, but the traditional UNIX interface doesn’t go very far in using that lower-level substrate to provide a higher-level services to the user and the programmer.
Great subject matter for an article. For a more in depth study of why ideas fail, or succeed, read the book, “The Tipping Point”, which is another term for critical mass.
More and more so, it’s the MS monopoly that is shutting down software innovation. What’s the point of even starting an innovative project, when Microsoft, with their soon to be 60 BILLION in the bank, can clone your work and give it away free to a zillion desktops? Talk about pissing on your parade! Not to mention the stiffeling effect on new ideas. Add to that, millions of brainwashed users that consider any software not bearing the windows logo not worth using. The built in obsolesence requiring repurchasing of applications one already owns, I guess is something that only bothers me.
The UNIX haters’ handbook seems like a manifesto for an alien set of mores for users; only it doesn’t roll them together as any kind of cohesive notion. That’s the punchline. There’s no rockabilly or SID music that rolls it all together in one bunch of SGI vector-movies.
Time required for something to catch on can go dry-hump itself; if something catches on, it gets fans and a connection to the world of leading ideas–dating simulation here, fiscal traceability there, dietary balance 3 li east of here. Sometimes, as they say, it takes 2 ideas.
I expected something like the book http://www.theburdenofbadideas.org/ from this article.
…whose thesis is that bad ideas are contrawise a resounding success! er… Or how about that there are 10,000 ways to take a break from logicism? http://www.manhattan-institute.org/html/_ws-bad_ideas.htm
Thumbs up on the LISP (Dylan) rant. I can hardly wait to buy the ARM implementation.