Slashdot Mirror
Lessons Proprietary Software Can Teach Open Source
cdlu writes "Kris Shaffer at Newsforge argues that just because software is open source doesn't mean it should be unpopular. What lessons, he asks, can open source projects learn from popular proprietary software?" From the article: "In the absence of a monopoly, there are three traits that are likely to make an application popular: it is cool or attractive in some way, it provides easy entry, and it is addictive. Barring these things, most average users will stick with the status quo. In fact, many users never use a program on their computer that did not come pre-installed. However, by creating an attractive, easy to set up, addictive application, a developer can motivate the average user to break this barrier and try something new. And several such applications can generate strong popular interest in the open source movement in general."
Any doubts about that? Check out the latest wave of Linux distros and their adoption rates. The distros that have live CDs are thriving. See Knoppix and Ubuntu for examples.
Agile Artisans
During my freshman year, I've watched a huge number of college kids switch to FireFox because of peer recommendations. Some of them even get OpenOffice.org and Thunderbird. I OSS software, especially for Windows, will continue to grow in popularity on quality alone.
The vast majority of closed source apps are sold, marketed. Partnering gets them on the desktop.
If we were talking about religions, closed source is Chrisianity, with missionaries, and wars and such.
Open source is Buhddism, where one must go and seek out enlightenment himself. There are no wars fought, to missionaries spreading the word. One adpots buhddism dur to principal, and not because someone else tried to sell it to me.
Appropriately, I think the world population of Christians vs Buhddists resembles that of closed-source vs open source. The same goes for adotion rates.
Slashdot's rate-of-post filter: Preventing you from posting too many great ideas at once.
That is frighteningly true. I made a program a while ago that tunnels a connection to another server while relaying the incoming stream to other users (a sort of MUD TV, called snoop, download it at www.poromenos.org), and I was amazed at the amount of questions I got about what I thought was self-explanatory. I ended up making an installation program with an option to install the settings for the MUD as default, because noone would use it otherwise (well, not without asking me dozens of questions about what the "remote server" should be).
Send email from the afterlife! Write your e-will at Dead Man's Switch.
Good FOSS projects seem to need more polished marketing. Firefox has made a good first step in this direction, but I have inevitably encountered resistance to adopting FOSS solutions in various workplaces, including small companies.
I'm not sure why this is, but when I show the decision makers a potential solution, the idea seems to be well-received until mentioning that it is free and open source, at which point interest seems to diminish. Recently, I was unable to get much consideration for pdfcreator, and it looks like we'll be buying a half dozen licenses of Acrobat, even though we just need each user to be able to generate a few (sometimes encrypted) pdfs each week.
I'm not sure why this is. Is there a perception of lower quality? A desire to have an official support channel (even though current support for most purchased software is atrocious)? Perhaps it's a mistaken, subconcious association between FOSS developers and hacking.
If it doesn't already exist, someone should set up a slick marketing website advocating FOSS solutions with materials for advocates to use in their workplace and content aimed toward purchasers who could use better education regarding what FOSS can provide.
Well, I think that for any piece of software to be popular:
It must provide functionality that is useful/interesting/fun (Productivity/Information/Games)
It must be easy to use, intuitive and of high enough quality that bugs are minimal
Software needs some form of advertising to make it popular. Popularity feeds popularity (Microsoft). Usually if the functionality offered is unique enough and useful enough, word of mouth/search engines take over and help with this.
Those people who think they know everything are a great annoyance to those of us who do.
Well, I wouldn't be using it, but we all know managers, PHBs and coworkers who would use it in a heartbeat.
Hell, I used to work with a network admin who played his favorite porn clips for general IT consumption, with the volume cranked loud enough for the customer service people upstairs to hear it!
If you want to make money off the CD, then start selling the links. Want links to your site included in PornLinux? Pony up the fee. Want some of your video clips (Complete with watermarks of course) on the CD in such a way as to make your site seem faster than the competitors? Pay the fee. Want to be a "preferred vendor" in the links on the CD, complete with links on the desktop instead of in the Bookmarks on the web browser? Pay the fee.
We all know Porn was the first thing on the Net to make money. Why should Linux be any different.
Now watch some bastard steal my idea, make a fortune and not pay me a royalty for the idea. I should patent it. Heaven knows the US Patent office would grant it in a heartbeat.
"Live Free or Die." Don't like it? Then keep out of the USA
This is an axe I've been meaning to grind for awhile now. OSS is like the world's biggest development and research laboratory. Given infinite resources and gallons and gallons of free code sloshing back and forth out there, OSS has yet to come up with something stellar.
That's not to say the OSS world hasn't made progress, and even come up with some interesting and useful things. I love it that I can open remote files over FTP from a KDE "open file" dialogue. I really love Jedit's plug-in architecture, not to mention its plug ins. I love auctex and emacs and save time with bash scripts and catalog my crap with a Mysql database.
So where's the radical new approach to software? I'm off to buy a copy of OS X Tiger because I want spotlight and dashboard for my Mac, knowing full well I can download Beagle and zeroconf for Linux.
I'm afraid all of the "but Windows users won't go for it" mentality is damping the creative juices of developers who are afraid to radically alter the computing paradigm in fear of alienating the Windows sheep that won't switch to any OS that doesn't exactly mimic the Windows software they use mediocrely. So we're forced to shoot for the lowest common denominator.
What would happen if, just for a moment, a group of smart people with full access to OSS code and no particular interest in pandering to the sheep put their minds together and came up with something radical?
I don't know what that radical thing would be -- I'm not one of those smart people -- but I do know computing is remarkably unchanged compared to the state of things 10 years ago. Linux has caught up with Windows as far as I can tell. So where is the innovation? What could we do if we weren't so busy trying to keep up with the boring monolith in Redmond?
If this were Usenet, I'd killfile the lot of you.
Just like anything that's already said, this one seems quite obvious.
.fla instead so someone could edit the presentation and make it, well, flashier.
Any FOSS product will be popular if:
- It is easy or easier to use than alternatives
- It gets the job done
- It gives something alternatives don't
- It provides as little as possible disruption
I would like to point out a couple examples:
I use Gaim on Windows XP (and under Linux - under OSX I prefer AdiumX, which is libgaim-based anyway) all the time. I have converted some people to it, but most of the non-conversions are due to lacking features like video or voice (I know it will be solved soon, if not already). It gives something MSN, Yahoo, ICQ and AIM don't: having more than one account logged on at the same time. Lacking features, tough, limit adoption. Running under Windows is a must - anything else limits adoption to, at most, 10% of the market.
My girlfriend was sold on Firefox because of the tabbed browsing. RSS is great and being able to import bookmarks is very convenient (But I am not very happy to lose the standard RSS links when I do so)
Similarly, OpenOffice.org Calc could win some users if it did something Excel would not do, like Monte Carlo analysis (I would love this one) or more than 256 columns on a single sheet (A client of mine would have switched from Excel just because of this). As it is, OOo Calc does neither. As a whole, OOo not being able to run natively under MacOS's GUI is also a problem.
I love to be able to export OOo Impress presentations as Flash movies, but I would like to add, forgive-me, more flashy features, like animated transitions. I would be very happy if I could export it as
Please note that ease of use means "it's easy to make it do what I want it to". Apache may be devilishly hard to use by a casual user, but a trained professional can make it do things IIS cannot, will not and would not even dare to try.
Well. My US$ 0.02...
http://www.dieblinkenlights.com
As always, quality is undefined. What kind of quality are we looking at here?
Some apps are rock solid while looking like hell being insanely difficult to install. That's not quality for me, even though you don't mind the looks and find the install easy (perhaps cause you've done it 15 times, getting the hang of it at your fifth time?).
Most if not all FOSS software are rock solid but are sadly lacking at 2 and 3. And that's what this is all about.
Neither Firefox nor OO or even Thunderbird have these problems. They're easy to install, run well and look decent. I'd guess all three factors came into play when your college kids decided to get and keep them.
Most proprietary software is rigorously tested on the lamen to see how well he/she can negotiate around it. Where as all but the most popular open source projects, frankly, don't give a shit.
While I agree generally with the thrust of your argument I think it may go a little to far. I do think many open source folks care about the interface. They just aren't very good at it and lack the resources. Serious interface testing requires a lot of resources that many open source projects find difficult to come by. They need to be able to observe how people use the product and that's not always easy.
I do think there is an opportunity for someone to create some open-source tools to help open source (and closed) with interface testing. (Maybe this exists, I'm just not aware of it) Imagine a tool which essentially records (screen capture) movies of users conducting certain tasks and also provides statistical data about things like time between button clicks, which menus were looked at and for how long, etc. I'm thinking something along the lines of a set of debugging tools (vaguely similar to a profiler I guess but for actions instead of code) which are oriented towards user interface work. The results could then be sent back to the programmers similar to how Mozilla uses TalkBack. This would solve at least one of the problems open source projects have in getting information about user interface problems.
Of course that doesn't mean the programmers will necessarily do anything with the data but at least it provides a method for those who take interfaces seriously to get some data to improve theirs.
am i really the first to mention it? AutoPackage should make things better for linux.. once users see some Click-Install action, they'll love it. (Personally I don't have a problem with Synaptic, but it's not what users are used to. I watched my friend using OS X once and he downloaded an app, and installed it without even thinking. Drag-dropped it right into the dockbar and he went and used it. Users tend to prefer this than starting up a special "install new software" app..)
I think a great analogy for this is the automotive industry: the people that design and build the engine are not the same people that design the dashboard/body/etc. The software created by the OS community are great engines. That's it.
While user testing is the best way to develop user friendly apps, there are known values and 'best practices' available to GUI designers that the hard core coder is not familiar with. Millions of dollars worth of university research is poured into understanding users and a lot of that info is freely available. Just using the basics can already improve many apps out there.
So, 2 things need to happen: 1- the OSS community needs to breed/recruit designers with a background in UI development. 2- Integration of the code and the UI needs to be easy to prototype and finish. As a designer, I know layout, but I don't know anything about windowing or developing in APIs. So I would need another piece of software (like VB or at the least the Design View of Access) where I can move around the widgets and components and graphics then mesh it all together later.