All Source Code Should Be Open

Developer.com ^ | 27 August 2002 | Charles Connel

[ShadowAce]

Penguin Ping

[Izzy Dunne]

How do we get software designers and programmers to raise the quality of their work?

How do you get furniture makers to raise the quality of their work?
By demanding it, or shopping somewhere else.
Think different.

[isthisnickcool]

He may have a point. But then again, he may not.

Copyright (C) 2002 - isthisnickcool

[rdb3]

The Penguin Ping.

Want on or off? Just holla!

[Apple Pan Dowdy]

"What about intellectual property rights? Wouldn't software companies engage in widespread theft of each other's hard work? No, they wouldn't, because this is already covered by copyright law.">

Copyright laws would prevent someone from reusing the EXACT same code, but wouldn't a few minor changes here and there constitute a whole new program? Example: a famous chef has his own recipe for grilled salmon. He will not publish the recipe because other chefs would use it and simply substitute or add one ingredient and thus have a "new" recipe.

Complicated programs are made up of many smaller bits of code. Would the copyright protect each of these smaller bits of code individually? Or could they be taken and used elsewhere?

[gilor]

More open source bullsh!t.

[Joe Bonforte]

When I bought a car recently, I didn't have to look at the engineering design specs or tear apart the engine to decide which one to get. I test drove the cars to get my own impressions, and checked information widely available in the market on the pros and cons of each car.

Why should software be any different? I'm an application software developer by trade, but I don't have the time or inclination to go through the source code for SQL Server. It's enough that it always does exactly what I want, gives great performance, and (if properly configured and maintained) is extremely reliable. Even if I went through that source code, I could never tell as much about the software as I can from using it firsthand.

So I don't give a flip if Microsoft releases the source code to SQL Server and most of their other products. Some of them are of questionable quality, and I don't buy or use those. Others are excellent and I use and recommend them.

They respond by continuing to improve the good ones and junking or revamping the bad ones. That's the free market. I'm happy with it.

Are you advocating force of law to make software developers open up their source code? Or are you merely suggesting that they "ought" to do so? I reread the article looking to see which you are advocating and it is not obvious to me.

If the latter ("all source code ought to be open"), then I don't agree with your opinion, but I don't see anything particularly egregious about it. If the former ("vendors should be required by law to open their source code"), then you are no friend of liberty.

[ShadowAce]

I prefer to liken code to literature. Each book is copyrighted, yet is composed of many words. Each word is obviously not copyrighted--merely their organization. As a result, we have great classics, and horrible bombs--but by studying examples of each, we can possibly create better books.

By opening the source of software, we (as a society) can examine, critique, and improve it. However, the author would still retain all ownership, with all the commercial benefits that entails.

[Born to Conserve]

So, they don't want us to patent the software, and they don't want us to keep the source secret.

Who are they and what do they want?

I sense a plot to redistribute wealth here.

[Illbay]

All Source Code Should Be Open...

And we will have the Party Central Committee FORCE those who continue to resist this universal truth. If their source isn't open, let them be sure that the Gulags ARE.

[ShadowAce]

I don't advocate law to require vendors to open their source. I'm not sure about the author's opinion, though.

I test drove the cars to get my own impressions, and checked information widely available in the market on the pros and cons of each car.

Why do you think there is information widely available? Cars are, essentially, open-source. If they were welded shut, with no access to the internals, how could you trust a test-drive or two? How would you know how well they are built? They might last for five years then just fall apart.

In that vein, you, personally, wouldn't be requireed to look through the source code for SQL Server, but you would know others have. And if no one raises a fuss over it, or if they praise it, you would know from an independant source that it was a good product.

[Illbay]

When you drive across a bridge, its design is open for inspection.

Oh, bullsh*t. Now he's talking about something that I, as a bridge engineer for some years, know about.

The fact is that very little of a bridge's "health" can be readily discerned by the casual observer.

The bridge in Oklahoma that collapsed earlier this year, did so only months after an inspection by trained inspectors. If he wants to use some kind of analogy, this ain't it.

Oh, and I have no idea how my CD player works. I squint a bit, and I can see some doohickeys and thingamabobs in there, some of which appear to go round and round, but I couldn't tell you if the thing's working correctly or about to give up the ghost.

[ShadowAce]

So, they don't want us to patent the software, ...

Note that I am not advocating open source licensing for commercial software.

[ShadowAce]

The fact is that very little of a bridge's "health" can be readily discerned by the casual observer.

And, as someone who knows very little of bridges, even I knew that one was weak.

[rdb3]

More open source bullsh!t.

Couldn't be that much. You responded, afterall.

[rdb3]

Why should software be any different?

Cars are cars. Software is software.

The analogy doesn't hold up.

[rdb3]

Note that I am not advocating open source licensing for commercial software.

No one on the OSS side is asking for leaving commercial software code wide open. Isn't it amazing how the detractors will misconstrue the argument anyway?

[jammer]

As a developer of a [large] niche product, I agree with the thrust of the linked articles. It seems to me that no matter how well-planned and written software may be, developments in most industries preclude knowing what changes and additions will be required in the future. Sure, OOP may help, but basic design of databases is just difficult to maintain in a changing world. For that reason, after kludges have been made, etc., for about 3-4 years, we attempt to rewrite the software. Of course, much of the previous work is used, but, until we can predict the future, this is, IMHO, necessary.

Of course, this expense is onerous, not only for designing and writing, but for testing and deployment. We have tried changing our philosophy more and more away from the "give the user what he wants at all costs" to "give the user a bulletproof tool that works WELL." Feature bloat is a problem (see the "office suites" whose every addition is probably useless and only adds to unreliability and user confusion), especially if a competitor says, "I can do this and Jammer cannot!"

However, I have a problem with distributing source code. The ultimate test is whether the software is adequately tested before deployment to ensure reliability and with enough error trapping (both user and program) to prevent disaster. Other than that, how can a user evaluate software? I cannot evaluate Windows source code. But, I can tell something is sucking wind when, for example, Win2K has to be reloaded when it capriciously loses the administrator and user passwords.

Constructive negative comments are welcome.

[gilor]

Couldn't be that much. You responded, afterall.

Yep, it's a fair cop. You got me there.