Slashdot Mirror
Big Ball Of Mud Development Model
Lightborn writes: "The Big Ball of Mud Development Model examines exactly why so many projects (software and otherwise) end up looking like a bowl of spaghetti. A good list of things not to do when developing a project."
This is a perfect example of why some code is crappy. One thing that always bothers me about Linux code in general is the lack of proper documentation. Sure, someone can say "read the f**king code, but in reality, it is usually much easier, even for experienced programmers, to have code that is well documented. If nothing else, it makes searching the code much easier.
In all of the commercial settings I have worked in, we always document our code. Each function is required to have a comment describing its function.
As a highly experienced developer of device drivers, embedded code, and other low-level code, the documentation is a big plus. Let's face it, computer code was designed to be parsed and read by computers, not human beings. One person's coding style may be hard to understand by even an experienced programmer.
The Linux kernel code is improving in this respect.
This post is encrypted twice with ROT-13. Documenting or attempting to crack this encryption is illegal.
Actually it's not an OSS problem. Bad software is generally caused by:
1) Bad initial design,
2) Bad/Inexperienced programmers,
3) Bad management.
OSS suffers from the first two, but peer review tends to weed out bad programmers pretty fast (at least in larger projects, less so in smaller ones).
Commercial software suffers from all three. The pressure from management to meet some imaginary 'deadline' (often invented after two many beers during an 'important meeting') means software goes out barely tested if at all. In 10 years as a programmer I've never seen a program get more than an hours testing before it got sent out.
Tony
exactly... people who want something that works, should use something that's LABELLED as working.
Hence, if you want a functional linux system, use 2.2.x, not 2.3.x
I'm sure things tend to break in FreeBSD-CURRENT as much as they do in Linux 2.(odd)
You are, of course, correct. However, OSS doesnt suffer as much as the closed source community from time constraints. Sure, OSS maintainers are harrassed by users for new versions and bug fixes, but in general, the maintainer's livelihood doesn't depend on releasing a new version. This gives OSS, in it's current form, a much bigger chance to produce better maintained code that most closed source software, as long as the oportunity to do so is utilized.
;)
Mind you, maintainers will probably have to put up with a lot of crap from users while they reorganize and reimpliment
Open source is massively contributing to better source, if it were only because more students now not only get the opportunity to read real-life code.
Even stronger, learning to read code has become more important than churning out vast amounts by yourself.
Inadvertedly, open source addresses a serious flaw in computer science education: the fact that students should, in the first place, learn to read, use, alter and inspire themselves from existing bodies of code, instead of churning out unrealisticly small mickey mouse examples.
Well-written code reads like poetry.
Can we be honest? The vast majority of it is complete crap, developed by amateurs with absolutely no clue how develop to professional standards.
I've only been working in the industry for a couple of years, and have interned at two companies and worked fulltime for a third. The code there was written by 'professionals' (in the sense that writing this code was people's professions) and has been varying degrees of ugly. Not always hideous, rarely fantastic.
I think that the Open source process encourages clean code more than closed source development, for two reasons. 1) You don't want to show the world your ugly dumb code, 2) the world has the chance to clean up your ugly dumb code, if it wants.
Trees can't go dancing
So do them a big favor
Pretend dancing stinks!
is a strong maintainer. A maintainer that rejects bad patches with extreme predjudice.
Ever read Linus' posts to Linux-Kernel? He is exactly the sort of maintainer, that calls out programmers that repeatedly submit incorrect or kludgey patches, and plants a boot up their backside right in front of everyone.
Read Linus, and you'll see he has a clear vision as to how the linux-kernel is supposed to work -- not just in a design capacity, but with an eye towards maintainability.
As Linus once told one of the GGI loonies: "the kernel isnt stable because it's a kernel, it's stable because I dont listen to people like you"
However, I have been using pair programming for about 2.5 years now, but only for hobby coding, and I have two things to say about it.
Trees can't go dancing
So do them a big favor
Pretend dancing stinks!
I've finally adopted a very game-programmer oriented philosophy towards development. Code should be written so that it is the specification, with appropriate inline comments documenting it and really clear variable names. Programmers should be extremely vigilant, and continuously roam their own code making sure that it actually reflects the current state of assumptions about the system. Whenever a change is made to the system, anything remotely affected should be proactively rewritten to reflect the change. This is pretty much how Abrash describes himself and Carmack working on Doom and Quake, and it is really successful. You keep performance up, stay in touch with your code, and never accumulate cruft. Bugs are immediately ferreted out and the programmer must never fear diving into code to tackle a big cleanup job, and can never allow pieces of code to exist that she (or he) doesn't understand.
Of course, you need massive automated tests to make sure your rewrites don't screw anything up. Designs must be extremely abstraction oriented, with a close eye to strong interfaces and bootstrapping, otherwise you will end up with so much code that it is impossible to manage the continual cleaning. And you need really dedicated programmers.
When I look at the Doom and Quake source, and the code that my own dev. team has produced, I see that the results are worthwhile. Each routine is beautifully crafted and works flawlessly. The codebase is a fraction of the size you would expect because so much effort has been put into doing everything the right way and eliminating broken or excessive code. And no bugs...
magic
I think this is an unfair generalisation. The vast majority of code I work with is clean, well-designed, and adheres to one of the style and coding standards out there. That includes the kernel sources, Glib, Gtk+, the Gnome libs, python, Perl, and loads of others.
If you mean that the vast majority of small applications are complete crap, you could be right. But don't knock the single programmer who need to scratch an itch. If enough people are interested in his or her project, they'll help - including design and style improvements and suggestions.
The OSS community needs to establish some quality standards. Linux code is relatively new, but this is going to bite everyone in the butt as the code gets modified more and more, and software rot starts to rear its ugly head.
Seems to be doing just fine so far - the OSS community is pretty aggressive about coding standards already.
Mark my words: Unless coding standards get real important soon, OSS is going to collapse under its own weight. "As long as it works" is not good enough.
Whereas of course closed-source software is bound to survive since all development houses rigidly enforce coding standards?
Bollocks. Been there, done that, shipped it because it compiled. In the bazaar model and peer review style of the OSS world, you can't get away with crappy code or bad design for very long. If your project's well designed, maintainable, useful and easy to read as well as being robust, it will survive. If not, then you can't get away with shipping only binaries to some poor customer.
--- Hot Shot City is particularly good.
Yep. However, I read some years ago that the Space Shuttle code costs $1000 per source line to develop (for the whole shebang, analysis, design, implementation, testing, maintenance, documentation, etc.) That's one thousand dollars per line (if I got paid that much, I'd be long retired :-). This only applies to manned missions though, software for unmanned missions costs about $100 per SLOC.
[cynical] I fear this is not because they value human life so much, it's more that loss of human life leads to huge costs in terms of publicity, scrutiny, congressional investigations, freezing of funding, halting of programs, etc. The Challenger tragedy cost far more than the $2 billion that the spacecraft costed, the halting of the launches and the redesign of the shuttle was far more expensive.[/cynical]
Anyway, as I should not have to point out, testing and bug finding is hampered by the law of diminishing returns. It is extremely costly to get the last 1% of the bugs out of the software. In the case of NASA and the space shuttle, there is an economic incentive to hunt down the last of the bugs. In the case of commercial software, stuff riddled with bugs is released because the cost of delaying the release outweighs the cost of leaving bugs in (dare I say that it often makes economic sense to leave bugs in? Just release the bug fixes as an upgrade...).
Anyhow, I think that Open Source has an advantage over commercial software in the sense that the developer(s) are motivated by something else than a paycheck. It is often a sense of pride that makes them strive for clean, bug-free code. Ever notice that the quality of software seems proportional to the inverse of the cost? I'm only half joking here....
MSN 8: Now Microsoft even has bugs in their ad campaigns.
I disagree. There is crap open-source stuff out there, but overall the coding standard is really not too bad, and some is excellent. The important stuff tends towards the excellent (at least from my experience).
By contrast, go and have a play with arbitrary pieces of Windows shareware . . .
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
12, 13 and 14 hour days are logged to make damn sure the spec is rock solid and ready to take over the world. Then it happened. The kind of comment that sounds like fingernails down the chalk board. The new VP (and CEOs son) stands up in the back of the room ans says:
" Hey! let's partner with Microsoft"
___
There's a pattern I've seen at some companies. Programmer comes in, spends 2 hours looking at the source. Says, "This is Crap, I need to re-write all this stuff." Then, he spends a long time, basically re-writing the entire thing, and breaking tons of stuff. At the end of a long process, the program is sort of where it was at the start, that programmer moves on and someone new comes in. He says "This is crap, I need to rewrite it all" and then the whole thing starts all over again. I would say that most programmers, coming into lot of 'someone else's code,' whatever the state of that code is, will tend to say it's crap no matter how it's put together. Just as a matter of reflex. And, partially, I think this is because they understand the stuff they wrote themselves, and when they read other code, they don't get the same warm fuzzy feeling they get reading their own code.
http://junglevision.com -- Shamus for Gameboy
This is somewhat similar to the softwear engineering methodology known as "Extreme Programming"
One of the main tenents of Extreme Programming is constant refactoring (ie, you see something that could be done better another way, you fix it straight away)
The other main point of extreme programming are: Always do the simplest thing that will work, and have proper, automatic test suites to constantly test your work.
Have a look at The Extreme Programming Web Site to learn more.
Pilot Light? Check!
One thing that gets me about the OSS community is the over-reliance on C.
Petrol(UK->US Translation - Gasoline)? Check!
I mean look at Gnome and GTK+, it based on some ugly C struct kludge to enable pseudo-object-orientation!
Flame-throwers are go!!
And then there is KDE 2.0 based on even uglier preprocessor commands.
I mean WTF is going on. The method of production in OSS is innovative, but the resulting programs end up being MS ripoffs. We need some true innovation regarding what we develop and the tools we use to do so. Because lets face it, most app level programming would really benefit from C++, or even something like Eiffel and the concepts of design by contract. Using pre/post conditions and invariants as in B notation, one can almost guarantee the correctness of one's program. (Eiffel and B were both in part developed by Betrand Meyer, he also played a hand in Z notation).
There is also an interesting project called EDMA That is trying to create an enviroment in which objects can be inherited from after they are built, a bit like CORBA, but IMO better.
void SelfPromotion {
I am in the early stages (i.e. thinking a lot and getting myself confused) of developing a Dynamic Object Enviroment to support reflection and better models of code reuse through selective "pilfering" of code and structure from other objects (Classes don't exist, only instances, although instances may share code). No links, or anything much to speak of as yet.
}
Anyway I digress, we should start thinking about the tools we are using, and ensure that they are suited to the job. For most things, the performance benefits of C, are not really crucial WRT anything outside the Kernel.
Cheers,
faichai
For a long time I couldn't figure out why others had such a hard time fixing bugs and changing their programs, while I could do it without any problems (no I'm not trying to be arrogant or pretentious).
Then a coworker made this remark: "Ray how can you write such well-organized code in one pass?" At first I couldn't understand what she was talking about, after all doesn't everybody constantly review their code? Doesn't everybody constantly rearrange functions and classes, rename variables, redefine protocols? Apparently not.
Correct me if I'm wrong, but it looks to me like most programmers write something once then spend the rest of the time trying to get that working. They never go back and rewrite the code, they just keep adding fixes to it. How can this ever work smoothly?
I've also seen and heard a lot about processes to make a program, or anything else, come out right the first time. I don't get that either! To me, the only objective when writing programs is to make it easy to change. Period. If it's easy change, it's easy to fix bugs, it's easy to enhance, and it's easy to rearrange and redesign with hindsight.
If you want to have a good time programming, do yourself a favour: learn the tools to make global changes to your code quickly, then spend a _lot_ of effort rearranging your code and renaming things as your program evolves.
OSS code looks worse only because you can see it. Commercial software houses have one big advantage, namely that they can pay people to slog through the ugly bits.
Unfortunately, non-open source also allows people to get away with crap they might be embarrassed to have included with their name in a publicly-available product. There's also a problem in that, except for the efforts a bunch of black-hats and white-hats with dissassemblers, most of us would never see just how bad some of the code in shipping products is.
In summary, I think it's extremely wrong to assume that OpenSource developers are more likely to be young, inexperienced and/or bad coders. I've found it to be more of a neutral as you have both good and bad programmers at work in either case, in roughly the same percentages and the ability of anyone to find bugs in OSS is countered to some degree by the QA departments for-profit non-open source companies fund. OpenSource has at least a small advantage in that the user can fix critical bugs, or would if they were capable of it and most are not.
__
I'm trying to convince my manager that this is the right approach for us but with little success.
I work for a small software house with just six developers on two main projects. What I see happening though is emergance of a structure that closely follows the XP guidelines. Our key developer left last month and left us with well over half a million lines of code to maintain and extend.
Naturally the first outcome was an outbreak of panic (particularly among the management) but us programmers didn't have time for lamenting. What happened though is that once we lost our mastermind of coding and had to rely on ourselves even with the most critical parts of the system people started sharing knowledge much more freely and the environment became extremely productive. I found myself often pairing with other developers to help them understand some parts of the code and vice versa.
It seems that with the old state of things (one developer churning out masses of code and the rest almost idle) while it felt comfortable it left many people fearing that they can't perform on their own (one of my colleagues wanted to resign straight away as soon as she learned that the guru is leaving).
The moral is that what pair programming prescribes turned out to be our life jacket and ultimately boosted the productivity of people who were previously intimidated by the presence of the 'main man'.
Personally I find coding in pairs a brilliant idea. I find myself producing much higher quality code when I have someone looking at what I'm typing. The bugs are fewer and more people know what's going on in the system. It works much better than peer reviews and documentation projects that we had foisted upon us by the management. It's not the official company policy but we do it anyway. For us XP works (or at least the parts we adopted).
This puts the research burdon on the person making the request. It also (hopefully) allows him to learn a few things during his quest to document his brain-dead demands thus lowering the likehood that we, as a company, have to waist time on a similar problem in the future.
___
Dun dun dun. Crash course in Managing Software(and hardware) Development. There are millions of links out there. So here's an introduction so people know what to look for. Remember, there is no magical model that automatically works!
e sting)->(Product life)
* -*(testing)->(product life)*[arrows back to design and analysis]
f ) -- more detail (http://www.incose.org/rwg/97panel/97 panel.html) - etc - (http://www.kingst on.ac.uk/~ma_s435/personal/work/CO1032B/tools_5/)
---
This is why you learn different management in software development models, because there is no one model that suits everyone. There are generally held principles that anyone can come to, but they aren't solutions that you can work out by common sense.
Here are some general statements:
"Software development is multi-dimensional."
OK, duh.
"Developers pay attention to what they are measured on"
OK, that makes sense. People _respect_ what management _inspects_.
"Some performance dimensions in software may be in conflict"
Makes sense, although a little more complex. (e.g., min memory, min SLOC vs. min effort and max user satisfaction vs max maintainability...)
Objectives in managing software development:
* define the process by which projects are conceived approved, and delivered
* define the guidelines and standards that are used by architects, developers and managers who will develop software
* define the mechanisms used to deliver the software to the marketplace
* general models to develop specific models in particular niche's such as "shrink wrapped" or "web based" or "b2b" or "b2c" or "OEM" etc
* define who is involved (e.g., product management, project management, development, technical writers, human factors/ui, localization etc) and their roles and their tasks.
* Specifications documents should follow these definitions and management models such as that for cost estimation (e.g., COCOMO, other models).
* once tasks are defined, you can help employees do what they are supposed to and evaluate them for future changes to development model
Interesting links:
a n article
The CMU software engineering institute
more
Defense system management college introduction to project management
wooha lots of links.
needed skepticism regarding empirical analysis with models!!!
"Commercial software models"
Example of cost estimation in use (findings from them at least):
http://www.ll.mit.edu/llrassp/jca/mcmb w.html
_Development models_ include (*== > in double sided->):
The incremental model;
AKA. The market model. Often dictated by management and generally follows QA builds.
(P.1)()()()()(1.0)()()()...(2.0)..
The evolutionary model;
AKA. The pseudo academic model
(Product Idea)*-*(Prototype)->(Clean Code)->(test and rinse)->(evolve)->(repeat)
The spiral model:
This model makes you ask the question as to the value of functionality and what process one would take in implementation.
(Kernel)->(Kernel+key or riskiest functionality)->(kernel+key+less troublesome components)->(K.+key+LTC+Less troublesome functionality)
Waterfall Model:
Intent:
(Product Idea)->(Analysis)->(Design)->(Implementation)->(t
Reality:
(PI)*-*(Analsysis)->(design)*-*(implementation)
Rapid Prototype model:
(product idea)->(prototype & analysis & design)->(implementation)->(testing)->(product life)
Common misuse:
(Product Idea)->(Prototype)->(More Code)->(Test)->(release)
etc, and hybrids like the "extreme programming" model, which seems to be a more detailed rapid prototype model
_Requirements methodologies_:
* generally: Requirements are what. Specifications are how (although they mix).
Incorrect requirements = no product, or bogus development plan
The method from which we develop requirements is:
discovery
refinement
modeling
specifications
requirements elicitation(href="http://www.se i.cmu.edu/pub/documents/92.reports/pdf/tr12.92.pd
How to defend against requirements crep:
* use formal methods !
* use customer requirements formats such as manuals or other docs !
* your answer must not always be yes !
* proposed changes must be evaluated and rational !
* there is always nearly a version 2.0 !
* the customer almost always values quality over a short delay !
* remain flexible enough to react to the work-place !
"without a manual, we don't have a product".
When you develop software by first designing it on paper with Functional Design and Technical Design reports, make database diagrams, N-Tier schema's where to put which functionality and code, UML schema's for your classes, you only have to type in what's already thought out. The only flaws you'll experience is bad algorithm coding (but that's not spagetti code) or detailbugs.
When you develop software using the evolutional model, that is: add code/functionality on the fly in an ongoing basis with short term designs and not based on original concepts and designs, you end up with eventually (most likely) a pile of code that has to be rewritten NOW because a new feature asks for it. because most of the time in these projects people do NOT choose to rewrite it, it's added anyway, resulting in spagetti.
In Short: evolutional model code is code where no theoretical basis is stated, there is no original manifest that illustrates WHY all the code is set up like this. MOST OSS projects are developped using the evolutional model. What helps is an ONGOING theoretical design document to function as a theoretical BASIS for the structure of the code. If there is NO designdocument or conceptdocument stating WHY code is structured the way it is structured, it's bad code. Period.
Another thing that adds up to bad coding is a bad naming scheme, or worse: no naming scheme at all! Nobody is forced to use hungarian coding, but please CHOOSE one! develop your own if no scheme suits you, but a scheme that HAS TO BE used by all developers in a project is a MUST to keep the code clean and updatable, even if you use designdocuments.
More and more OSS projects get tighter software teams with people who KNOW how to develop software, thus using designs and theory before starting to jam in the code, and that is a good thing. We ain't there yet however. For starters I'd suggest to take a look at the InfoZip sourcecode: ansi and non-ansi C together in 1 project... *UUHHHHH*
--
Never underestimate the relief of true separation of Religion and State.
I'm currently working on a game which, when it's playable, I intend to release under the GPL (I know, 'release early, release often', but still, I don't like releasing a nonfinished product) -- I find the idea that literally the entire community of linux users could be looking at my source code is an enormous incentive to make it clean.
I guess the same force that motivates me to clean up the living room when people come over also motivates me to clean up my code when people are going to look at it. It's enough that they'll see how naieve a coder I am, I don't want them to think I'm a bad coder too.
Which reminds me, I have to go do the dishes....
-Denor
I agree with you in part, but pigeonholing all of OO as obscurantist and overcomplicated is as erroneous as overhyping it as if it were the Second Coming. There is nothing wrong with the motivation behind OO itself; the problem lies mainly in implementation. Specifically, what we now consider to be "OO programming languages" and "OO design practices" goes far beyond the original concept of both, and much of it is indeed obscurantist nonsense, which induces a huge amount of needless overhead, both of the conceptual kind on the designer and of the practical kind on the implementer. This is especially true in the case of small to medium software projects, even more so because often designer and implementer are one and the same.
Case study 1: C++. An extensive critique of C++ as an OO language for production systems, from the point of view of an Eiffel-cheerer, can be found here; in my opinion, it suffices to say that, given its status as just one step up from a C add-on, and given that, when building on such a shoddy conceptual infrastructure as C's, it's hard to conceive how one could do any better, C++ should be considered to be outside the scope of this discussion.
Case study 2: Java. Now, Java is built from head to toe for maximum OO. This is incredibly intrusive to anyone who wants to do some real work using it, as opposed to just drawing nice schemes and writing UML models. Java is built to enforce those styles and concepts of programming which the designer felt to be correct. It's languages like this which give OO a bad name, and they should be shunned.
Case study 3: Perl. Perl was built to be a scripting language - in Osterhout's original conception, a "glue" language. Thus, practicality being the most important goal in it, it's easy to understand why Perl's OO is as it is. Specifically, it doesn't exist per se; no special syntax or semantics is enforced for OO programming, in fact all of it is built upon simpler, pre-existing constructs - specifically, taking advantage of an isomorphy between modules and classes, objects and references (via abuse of the bless() and ref() functions), methods and namespace-local subs. This makes a transition to OO practices easier as a project grows. It also allows one to implement the concept of an object as he sees fit - usually the slot approach is used, using hashrefs, but there other approaches for specialised cases - including objects as indices into class-wide property arrays, an approach described in "Advanced Programming with Perl" and which is useful for when you need many objects and creating a hash table for each would be a waste of space. The discussion of OO in Perl could be extended further, but it suffices to say that, in true Perl form, it restrains from imposing a paradigm on the programmer, trusting instead that he knows better.
Case study 4: Smalltalk. Smalltalk is widely considered to be the godfather of modern OO (yes, Modula had something called "OO" before Smalltalk, but a quick glance at both languages will make it clear right away that most of what we call OO today was fathered by Smalltalk); this, combined with the widespread availability of "OO software design tools" for the environment, could lead to some people blaming it unfairly for their current issues with the paradigm. In reality, when using Squeak, a computing environment integrated with a derivative of Smalltalk, I've found the use of OO in programming the system to be perfectly natural, in contrast to the uncomfortable feeling that you get from using, e.g., Java. Part of this comes from language design itself, which makes the concept ubiquitous in a very straightforward and graspable way, but most of it comes from the environment, which is fully built on objects. In the Morphic system, you can "see" and "touch" - inspect, manipulate, delete - all objects alike. The user- and programmer-levels are intertwined, and so instead of programs, methods are the elementary user-level executable unit; this removes one unnecessary level of encapsulation, leaving all objects free to talk to each other, without being first streamlined into the procedural mold enforced by the "program" concept. All of this, plus the elegance of the Smalltalk language, makes for a system which is very easy to program, and which leaves relatively little to be desired. Thus, I consider Squeak to be a paradigm of well-used OO.
Hell, I think I've said more than I set out to... I hope at least some of it is of any use.
To the editors: your English is as bad as your Perl. Please go back to grade school.
One of the good things about OO languages (and I'm not particularly fond of OO) is that they make you think about your data more. OO is not a silver bullet, though, since it's certainly possible to use one to organize your data badly. No language is a substitute for an experienced developer with some talent for organizing data in the right way for the particular project.
Of course, this is not a new concept. Fred Brooks said it nicely in The Mythical Man-Month, a book which should be required reading for everyone who does software development, and more so for people who manage development efforts.
Welcome to the real world of software development!
Just be sure to wear the gold uniform when you beam down -- you know what happens when you wear the red one.
I'm reminded of the Dilbert where the PHB says something similar and after praising his insight they completely ignore him and carry on doing it the same way as before.. just a thought.
Any sufficiently advanced man is indistinguishable from God
I'm really glad to see this. In my experience, the great flaw in the OSS model is the quality of the code. Can we be honest? The vast majority of it is complete crap, developed by amateurs with absolutely no clue how develop to professional standards.
The OSS community needs to establish some quality standards. Linux code is relatively new, but this is going to bite everyone in the butt as the code gets modified more and more, and software rot starts to rear its ugly head.
Unfortunately, the vast majority of OSS developers are not very old (less than 25), and don't have the perspective to appreciate trying to maintain 10 year old code that has been modified 20 zillion times.
Mark my words: Unless coding standards get real important soon, OSS is going to collapse under its own weight. "As long as it works" is not good enough.
--
Sometimes it's best to just let stupid people be stupid.
Why do I get the feeling this problem isn't just found in OpenSource projects? Zillions of programms, both free and commercial, are badly designed from the start. Many more could be well designed if only they didn't have to worry about backward compatibility. (probably one of the biggest problems for Windows right now...) The Big Ball of Mud architecture isn't uncommon by any means. And it's not a problem that only OpenSource faces.
The vast majority *is* crap. But the stuff that is important, libc, the Linux kernel, GCC etc. isn't. If you look at Windows there is lots of third party software that is complete junk. The case is the same with Linux. But the "big stuff" is all big *because* it's good, well designed software. OpenSource can produce crap, anyone armed with a compiler (standard on most UNIX's) can produce utter junk. But will that junk be used? Will anyone even know it exists? Of course not.
Both OpenSource and Closed Source development can produce junk software. And both can produce great software.
This is *not* just an OSS problem; it's a problem with *all* software. Even stuff which is well-designed from the beginning, and reasonably well-written, degenerates over time; and with high engineer turnover, even stuff written four or five years ago becomes painful to maintain.
For the OSS community *in isolation* to seek to solve this problem would be unfortunate; the industry as a whole needs to find a way to address it.
One example of some very well designed software is the Shuttle OS that powers NASA's Space Shuttle. In 420k lines of code each revision has only had 1 bug each that wasn't caught by testing. If we *really* want to make some good software it wouldn't be a bad idea to take these lessons to heart. OpenSource software is already good, lets make it better. Full artical here.
One of the emerging trends in academic "out there" approaches to software engineering has been moving patterns from architecture (sofware and housing) to behavioral practice, of which this is a negative example. For positive examples, there is an intesting website I ran across the other day about a concept called "Extreme Programming". This concept wouldn't scale to the sort of distributed development done for OSS, but portions of it might. In any case, documenting robust positive patterns for OSS development sounds like an interesting project.
I thought the comparisons to Design Patterns were hilarious but way too true.
Bleh!
Your impression of open source developers is distinctly wrong. You need to seperate the people love linux because they grew up with it from the people who actually do development and can make commits to the kernal. You are making a rather broad generalization when you say
"Unfortunately, the vast majority of OSS developers are not very old (less than 25), and don't have the perspective to appreciate trying to maintain 10 year old code that has been modified 20 zillion times."
Source? Your personal opinion? If you take a look at the people work actually make significant open source project work, they in no way match your description. When you say the vast majority do you mean the vast majority of OSS projects that get started and then dropped? That's the nature of OSS. Crap generally make it into the kernal or a mature project.
Thalasar
Anti Patterns: Refactoring Software, Archetectures and Projects in Crisis
ISBN 0-471-19713-0
Excellent book, I would highly recommend getting your bosses a copy or get a copy if you are running a project, its concise, incisive and useful, personally I'd rate it up there with The Mythical Man Month and helps when you need to point out the company is making a classic mistake.
Any sufficiently advanced man is indistinguishable from God
DocBook (and all other forms of inline documentation) suck ass. However, they are about 10 times better than no documentation, and 1000 times better than inaccurate documentation. Experience shows that using out-of-line documentation leads to one of the two above problems, and therefore auto generated documenation is becoming the standard.
So, in practice I agree with you, but there is a widely held belief that inline docs are inherently superior to out of line docs. This is just plain wrong, as it leads to suboptimal documentation and suboptimal code (I have often cursed wading through gobs of commented out inline documentation when all I really wanted to see was the code). This is one huge advantage (of many) of C/C++ header files over Java-- at least the documentation only obsures the declaration, not the actual code.
Large Scale C++ Software Design, by John Lakos. This book has done more to improve my coding and software design skills than any other book I have read. If you program in C++, you MUST read this book. Until you have, you don't know the language. The concepts described in the book apply to other languages as well (as long as you are using OOP).
------
Not long ago, there was a slashdot post about the Linux Kernel 2.4 to-do list.
Someone complained that the list was proof that linux kernel development, and open source development overall, is bad.
The argument was that any decent system would keep everything working all the time. My reply was that innovation doesn't come easily, and that you can't improve a system, while keeping all of its parts working the entire time.
It's no wonder that closed software gets so bad and bloated; they're all probably doing the very things listed in this ball of mud article. "Daily builds" can sound like such a good idea, but they do lead to problems.
Why does spaghetti end up all tangled up? Or, a more practical problem, why do power cords, cables, phone lines, etc. all end up in a gigantic ball underneath my desk, no matter how many times I sort them out? I don't think it is just Murphy's Law.
My own theory on this is that the cable, cord, etc. tends to be more weighted down in the center, and therefore drags there. As it drags, it forces other cables, cords, etc. out of the way. As soon as it forces them out of the way and drops down to a lower level, the tension on the other cords, cables, etc. tends to make them wrap around the heavier cables cords...so in about a month or so, they all tend to tie around each other.
Hopefully I didn't put any [] around my words.
Ball of mud? Bad ideas? Low quality?
No fair. I've been working on patenting these techniques.
---
Dammit, my mom is not a Karma whore!
When I write, I go along a line of argumentation until something starts feeling wrong, like I've strayed too far. That's when I go back, read all that I have written, fix it up, and then continue writing until I have to stop again.
When I code, I do exactly the same thing: code until it feels too messy, go back, rework, continue to code anew, get stuck, etc.
The result has been fairly decent code that isn't too bad to alter over time. However, sometimes I get tempted to overhaul code when it really isn't necessary, because some minor issues are bothering me. (This happened with GeekPress when I was just a few days of programming away from launch, but thankfully my husband helped me get over my fussiness!)
Since I've always completely coded my own projects (even when working within a company), I have no idea whether others code in a similar fashion or not. (I'm sure that my situation is greatly simplified by the fact that I don't have co-programmers. That seems like a nightmare to me!)
-- Diana Hsieh
-- Diana Hsieh
GeekPress: The Weirder Side of Tech News
Some of the more respectable C++ journals have recently done some good article and interviews about Extreme Programming. I highly recommend looking into it; even if you adopt none of its practices, the concepts raise good points.
You cannot apply a technological solution to a sociological problem. (Edwards' Law)
>It is hard to find really good Software Architects.
I figure that SW Architects are a bit like the other kind of Architect (those who design buildings), so it might be worth mentioning one kind of architect: those who Want To Be Artists, but decided they like a steady paycheck, don't want to starve to death in a garret, so they get a degree in architecture.
These architects are the most disorganized pains in the ass, always working to the last moment when they decide ``'we have to stop & leave the mistakes in", thus throwing off the timetable for deliverables & making the rest of the group pissed off. And they usually don't deliver very good work, either.
I figure there are numeorus members of this school of architecture working in Redmond right now.
Geoff
I think I see a trend here. Maybe for them it really would be easier to muzzle the entire internet than to produce p
WWJD? JWRTFM!!!
You can set your threshold to read -2 postings. Use the controls at the top of the comments listing to set your threshold to -1 and check the Save box, then click "Change". Then manually edit the Threshold field in the URL that gets returned so that it says "threshold=-2", then hit Enter. Voila!
--
--
Do I look like I speak for my employer?
Just the other day I was in talking with the lead developer for one of the projects I was on over the last 9 months, explaining to him how the last 7 years of code that had been written on the project was completely underdesigned and thrown together with duct tape and bailing wire. I cannot even begin to describe it. And I am supposed to be the buildmaster for this, and we are looking to have offsite developers go to town on this stuff. He looked at me, and listened to me bitch and then said "Well, we don't want to invest a lot of time into this, or money." WHat??!?!?!? He wants me to restructure the last seven years of work in a slap dash manner, not even fixing the problem, or making it worse! It was exactly this kind of thinking that got me to the point where I had to say this has got to stop. I may be just ranting, but what the hell else am I going to do? I am supposed to be buildmaster and tech support. Well, because it is such a bear to install, we have like 3 users. Guess what? I don't have anything to do but fix it! SO what am I going to do? I am going to come in on weekends and work when he is not looking and make it go, and do it right, because I am _offended_ by the crappiness of the hierarchy and code design. Now, I am not a total lone wolf here, I am discussing the structure with them, but people need to realize that a crappy design leads to a crappy project.
-- Who is the bigger fool? The fool or the fool who follows him? --
It is hard to find really good Software Architects. Lately the tendency has being to produce code at the speed of rabbit procreation - 10 times a week. The problem has gotten worse over time due to the popular rapid development tools such as Visual Basic for Windows and Object Oriented approaches such as Java for JVM. The bad news is that this deters new good programmers to appear instead of old ones. The old programming school did not have to face such issues as handling millions of users and huge databases or creating user interfaces accessible to a novice user, they were mostly concerned with the speed. Ability to hack together some brilliant Assembly code was the primary concern, I admit it is cool.
Today most so-called Microsoft Certified 'Engineers' have no clue what 'Assembly' stands for but they still don't know how to handle millions of users terrabytes of data or create decent user interfaces. The problem is that computer science became popular among those people who have no real call in their lives and who regard their work as simply a way of getting their salary. Large salaries of IT department does not help too much, they create an unhealthy attitude towards the profession.
Working on a large project that is supposed to be scalable to millions of customers, supposed to handle multiple user interfaces of various wireless devices (PDAs, Cell Phones etc) over time I had to design various components of the system. In the beginning there was only an idea which later became basically a large collection of various components. I have never before had to design and build such a complex piece of software and I am just happy that my current formal education allows me to make sound judgements about network traffic averages and variances, the speed of code in terms of iterations (big-O, big-Omega, big-Theta, they are usefull after all), being able to handle various datastructures and even creating my own new tree designs.
Nevertheless, all the way through I've felt the need for an experienced software architect. My company did not have one and we still don't and I think it is very difficult to find one with really good experience and skills.
Once I have seing a real software architect at work (he was in his forties) he was giving a presentation of his design and it was jus WOW. I mean even after working professionally for three years and handling hundreds of different programming and design problems, I don't think I could produced such a thoughtfull design that goes into details and goes over every possible issue with all the computations and considerations. It was beautiful.
I wish we all could learn from the best.
You can't handle the truth.