“Of all the myths that have grown up around open source software, perhaps the most pervasive is Eric Raymond’s aphorism that ‘Many eyes make bugs shallow’, suggesting that if lots of people can view a program’s source code, they will find and fix its errors more quickly than commercial products whose code is jealously guarded. The only problem with this is that it’s not true – certainly not in one of the flagship projects of open source, OpenOffice.”
While what the author claims may be true for applications such as Open Office, there are a great many projects out there that most definitely do benefit from being open source. Just look at something mission critical like the Linux kernel, or OpenSSH. It just makes sense that the many eyes approach is going to find more bugs in those products than if they weren’t open. And once they are found they certainly get fixed quite rapidly.
I wouldn’t exactly call the Linux kernel mission-critical … at least not 2.6.
Well, anyone but you, I would keep quiet.
Now, tell me how much did you test 2.6 and why not mission critical, or better what is mission critical for you? I’ve got about 60 servers (not even counting desktops into this reference) running on 2.6 (started using 2.6 after 2.6.3) and without a single flaw.
Can’t say that about my one and only Windows server deployment. One OS crash and two disk failures (partition not hdd) in one year and a half.
There is just one rule to this game, you must buy 100% Linux certified machine. And no experimental kernel drivers. Just few sample cases for basic server setup.
Opteron yes, Xeon no.
NVidia yes, Ati no.
Adaptec RAID yes, LSI no.
SMC yes, Broadcom no.
real (and supported) modem yes, winmodem no (you need that if you need to set up fax gateway).
etc… but still, you have to be carefull when you buy Windows server too.
Just looking at the hardware I buy was reason enough to avoid any proprietary driver. Always directly and completely supported from setup.
Hell, even my notebook worked like that. Again can’t say that about Windows.
Based on inteligence of your usual comments, I’d say you have tried few distros for a day or two and didn’t liked them, because it just wasn’t your world. So you just packed up your toys, left the kindergarten sandbox and started promoting stupidity.
Yeah, of course … you guessed me right to the dot! Good job.
Anyway, I blame your Windows experience on you and you alone. There are many 2000/Server 2003 server deployments out there, and as far as I know, they’re running just as smoothly as their *nix counterparts.
The 2.6 kernel is a developer’s playground. Rather than just fixing bugs, they pack more features in with every release. You can’t reduce the number of bugs when you’re packing features in. Do I have to quote Ted T’so on what he thinks of the 2.6 kernel? Don’t make me.
That is probably why I only use the latest 2.4 linux kernel version =).
However with each 2.6 iteration it gets tighter and the code gets cleaned up. I will only upgrade to 2.6 when the developers move on to version 2.8.
Of course 2.6 is a developers playground, but to say that they are just not fixing bugs and implementing new features is not entirely true. The kernel will reach a point where the code is cleaned up, and the feature list is standardized.
In fact the way I look at things. 2.6 is beta until they move on to 2.8. Seeing as how minor upgrades happen so fast and so often (what we on 2.6.13 right now?). And Im fine with that. So you can quote whoever on whatever. Im currently impressed with how things are going.
Using a incremental design and implementation/ protoyping/ rapid development method for software (via linux) is different from having the well known waterfall and deployment method (via windows). But it is still proven and viable. So meh.
Sorry linux isn’t what you want it to be. But that doesnt warrent any reason for you to think that development is just stagnant. Same goes for Ted T’so.
Ted T’so is a senior Linux kernel developer.
“Not all 2.6.x kernels will be good; but if we do releases every 1 or 2 weeks, some of them *will* be good.”
Great professional attitude there. “Hey, we don’t know which ones will be good, so if we release a bunch one of them is bound to work right!”
That is why I don’t touch 2.6.
And you’re taking that completely out of context from an extremely long thread debating the various ways that kernel devs should go about doing releases. If you’d even so much as look at the full text of Ted’s post you’d see that he’s merely considering an alternative to doing release candidates before each release.
But granted the 2.6 kernel series has seen some very heavy development. So if you’re not up to dealing with the hazards that may come from running the latest and greatest kernel then stick with something stable that’s been packaged and tested by your distro of choice.
While new vanilla kernels are released every now and again, this doesn’t mean that every time a new one pops up everyone is forced to upgrade to it. And anyway, Linus has publicly stated that with 2.6 kernels it’s up to the distros to make sure they’re stable. So just to give one example, Red Hat has been going with the 2.6.9 kernel for quite a while now, and I expect they’ll stick with it. And it has been quite stable on my end. I’d also expect similar things going on with Suse and Debian and others.
But then I don’t know why in the world I’m even responding to you when all you’re really after is getting a rise out of people. Oh well, maybe you’ll grow up one day kid.
Yeah, of course … you guessed me right to the dot! Good job.
But you did a bad job on mine
Anyway, I blame your Windows experience on you and you alone.
Funny you say that. You don’t know what the problem was and you blame me:) Another one of your childish ignorant conclusions (or better said, delusions).
Involved with Windows from 3.0. Coding system services mostly (and just for the fact, most of my MCSE friends ask me for help not the other way around). I could fill up a resume, but I really don’t see the need because you haven’t said anything constructive enough to deserve that.
they’re running just as smoothly as their *nix counterparts.
As smooth or as long? It is a difference. And this could be another topic, why I left the closed world after being in there for so long. Again you haven’t had one constructive thought for me to feel the need to explain you. That would be taken in account all comments I was reading from you, not just this one.
The 2.6 kernel is a developer’s playground. Rather than just fixing bugs, they pack more features in with every release.
Ok, here is a thought. If 2.6 is developers playground, then Windows world is a battlefield. One side MS with their vanilla kernels (which don’t support by default one single machine (without installing additional HW drivers) you can buy now) and HW and SW vendors on the other side.
It is a world where MS doesn’t test their service packs with all drivers on this world and where vendors don’t test all service packs.
Which is better? 2.6 comes in one package, at least meaning it was tested like that. Windows is never tested like that.
Most of the HW vendors don’t have or don’t care to access the Windows sources and all the need they feel is that they produce drivers that somehow work.
If you think that HW vendors make perfect drivers, then just look at any driver changelog, or BIOS version changelog. They do make mistake just as anybody.
You can’t reduce the number of bugs when you’re packing features in. Do I have to quote Ted T’so on what he thinks of the 2.6 kernel?
And how could you reduce number of bugs if you don’t even make the drivers, or you make drivers and you’re not involved with OS. It i hard to run for olimpics when you only have one leg.
Don’t make me.
What? Childish,… stupid,… ignorant,… delusional? People don’t need to make you that. You took care of that already.
p.s. You’re doing a good job in that. Image you present, I mean.
> Funny you say that. You don’t know what the problem was and you blame me:) Another one of your childish ignorant conclusions (or better said, delusions).
If you used Windows, and Windows broke for you in serious ways, while others use Windows and it works fine, it’s safe to blame you. You think Microsoft’s NTFS code isn’t tested? Why else would the partition become corrupted? Bad hardware? Bad sysadmining?
> As smooth or as long? It is a difference. And this could be another topic, why I left the closed world after being in there for so long. Again you haven’t had one constructive thought for me to feel the need to explain you. That would be taken in account all comments I was reading from you, not just this one.
Smoothly, as in not causing any headaches. Long too, I suppose. There’s no reason why an NT box can’t stay up for years — many have.
> Ok, here is a thought. If 2.6 is developers playground, then Windows world is a battlefield. One side MS with their vanilla kernels (which don’t support by default one single machine (without installing additional HW drivers) you can buy now) and HW and SW vendors on the other side.
So you expect Microsoft to develop drivers for every single piece of hardware out there? Are you insane, or just stupid? The original manufacturer can do a better job of developing drivers for their own hardware — most of the time. The NT kernel’s job is not to support hardware out of the box — Microsoft provides a set of drivers for common hardware in Windows, but nothing for hardware that came out after the release.
> It is a world where MS doesn’t test their service packs with all drivers on this world and where vendors don’t test all service packs.
Great expectations for Microsoft, excuses for Linux. Can you honestly say that devs test their shit on every single hardware combination possible? Nope. Microsoft already does extensive testing of many things, but testing SPs with every driver in existence is a technical impossibility if the SP is to come out in due time.
> Which is better? 2.6 comes in one package, at least meaning it was tested like that. Windows is never tested like that.
You know this? No, you don’t. Go read any number of articles out there on the Windows development process, and you’ll know that they have testing infrastructures in place that kernel developers could only dream of. Each daily build of Windows goes through hundreds of thousands of rigorous tests at night, with full reports the next morning.
Just because 2.6 is released as a “single package” doesn’t mean it was tested in its entirety. How many times have we seen bugs introduced somewhere as a result of a minor change in some driver or other? Or regressions in performance/stability? Get a frigging clue and stop being such a Linux fanboy.
> Most of the HW vendors don’t have or don’t care to access the Windows sources and all the need they feel is that they produce drivers that somehow work.
Again, you know this how? Quit making unfounded statements. Either back it up or don’t make the statement. Many hardware manufacturers strive to produce great drivers. NVIDIA, ATI, and VIA are just a few to name.
> If you think that HW vendors make perfect drivers, then just look at any driver changelog, or BIOS version changelog. They do make mistake just as anybody.
I never said anything was perfect. I realize that drivers have bugs. I realize that BIOSes have bugs. I realize that Windows has bugs. I realize that Linux also has bugs. Your stance on things seems to be that Linux’s bugs are all either minor or non-existant, while everything is flawed.
> And how could you reduce number of bugs if you don’t even make the drivers, or you make drivers and you’re not involved with OS. It i hard to run for olimpics when you only have one leg.
Well how about this — rather than packing in features, how about fix the ones already there? Hey, there’s a great idea!
And no, you don’t have to be involved with OS development to create drivers. The hundreds of thousands of drivers for all the versions of Windows out there are a testament to that fact. You lose.
I guess throwing out insults and unfounded statements makes you feel like a big man. Do you put that on your resume too?
“Able to throw out vitriol and unfounded statements on a whim.”
If you used Windows, and Windows broke for you in serious ways, while others use Windows and it works fine, it’s safe to blame you. You think Microsoft’s NTFS code isn’t tested? Why else would the partition become corrupted? Bad hardware? Bad sysadmining?
Nope, commercial app on Exchange. After Exchange SP was applied (and yes, devs said it is safe to apply SP, I checked that with devs before applying SP). Is that bad sysadmining? Now, I expect you’ll say bad application or bad developers. Remember, it is me who is the biased one, at least you say that after a while. We will get there. So you can’t afford this in your response, remember you’re non-biased one. So app colapsed server, which colapsed fs. Whose fault it is? For sure it wasn’t hardware fault. In any case Exchange and NTFS shouldn’t colapse just because one app crashed
Smoothly, as in not causing any headaches. Long too, I suppose. There’s no reason why an NT box can’t stay up for years — many have.
Yep, said that too. I was arguing about smooth:)
So you expect Microsoft to develop drivers for every single piece of hardware out there? Are you insane, or just stupid? The original manufacturer can do a better job of developing drivers for their own hardware — most of the time. The NT kernel’s job is not to support hardware out of the box — Microsoft provides a set of drivers for common hardware in Windows, but nothing for hardware that came out after the release.
Nope, but I remember that in lot of your comments you expect that from Linux devs:) That would be your comments about bad HW support. I guess that makes you the crazy one. It isn’t me who is insulting you, it is your history.
Great expectations for Microsoft, excuses for Linux. Can you honestly say that devs test their shit on every single hardware combination possible? Nope. Microsoft already does extensive testing of many things, but testing SPs with every driver in existence is a technical impossibility if the SP is to come out in due time.
So saying the fact that Linux includes number of drivers while Windows supports only older in its vanilla state is biased approach to commenting.
Remember, just as you might be bugged with Linux HW support, I don’t like to go trough driver, restart, driver, restart just to set up my system (and here it only starts). It is so much easier to buy certified machine pop a cd and enjoy.
So yes, you could call it, my published annoyance with Windows. And just one of my many reasons that Windows is avoided as much as possible.
You know this? No, you don’t. Go read any number of articles out there on the Windows development process, and you’ll know that they have testing infrastructures in place that kernel developers could only dream of. Each daily build of Windows goes through hundreds of thousands of rigorous tests at night, with full reports the next morning.
So this is why SP2 broke so many apps:) Even MS aknowledged that before putting SP out. Most of that aknowledgement was taken from beta SP2 where users bugged MS, not by MS alone
Again, you know this how? Quit making unfounded statements. Either back it up or don’t make the statement. Many hardware manufacturers strive to produce great drivers. NVIDIA, ATI, and VIA are just a few to name.
NVidia, agreed. They do make great drivers, at least I can’t remember having one single problem with them, be it on Linux or Windows. Neither I can remember having them with intel for their MBoards from LX440 to 845 (815 is exception in that time. There were just too many variants of this board, but if original CD was at hand no problem, except with the sound card few times). Problems with intel started after 845
Other two I’m just laughing. My experience with computers says one thing only. Stay away from that half assed hardware.
Next example would be Matrox. It was all good and glorious until Parphelia. .NET driver just sucked major.
I take care for about 60 servers and 250 desktops (and there is a 4 year computer lifetime policy). Enough testing ground?
p.s. Again, you expect backing up their comments from others. Why don’t you meet your expectations:) If you would then you wouldn’t talk about things you don’t have a clue about, just the last example. Comparing *X “find” and Windows “find”. Well, that one was a real eye opener. Can disect your comments if you’d like
I never said anything was perfect. I realize that drivers have bugs. I realize that BIOSes have bugs. I realize that Windows has bugs. I realize that Linux also has bugs. Your stance on things seems to be that Linux’s bugs are all either minor or non-existant, while everything is flawed.
??? In my first comment I described that better. It is perfect in right conditions. And I even said when it is not perfect.
Well how about this — rather than packing in features, how about fix the ones already there? Hey, there’s a great idea!
Obviosly, as expected you’ve never read any changelog for linux
And no, you don’t have to be involved with OS development to create drivers. The hundreds of thousands of drivers for all the versions of Windows out there are a testament to that fact.
Yep, and you don’t need to have basic computer knowledge to write a clustered CS app.
p.s. in translation, “it requires deep knowledge of system internals to write driver, where best condition would be if you would write that OS”
You lose.
What? Did we compete?
Or did you actualy said something constructive and I’ve missed that?
I guess throwing out insults and unfounded statements makes you feel like a big man. Do you put that on your resume too?
Funny, “YOU” said that
Thanks for this one, best one of the day:)
p.s. In long comments you should use html tags to label C&R, it is hard to read with >
I’m not going to bother responding to this. It’s just badly-written anti-Microsoft FUD and Linux-loving.
You remind of Moullineuf, though not quite as bad.
And your comments would be good based on what?
1. You never stand behind with the facts (yet you always accuse others of doing that)
2. You don’t need much to become insulting (there is a lot of comments in your history that can be taken as proof, yet again accusing others doing that)
3. You’re writing could be described just as MS-loving, anti-Linux FUD (sadly, you don’t stand to back it up. Which would make it better for MS if you would keep quiet. At least learn the facts before you speak). Wow, that is easier to say than actualy backing facts. One can smell over the internet you don’t have a clue about Linux or *X. Just answer this. How can one compare two sides when he doesn’t know the other side. Again example with “find”. Any kid knows that “find” in Windows is a joke, while *X command actulay provides you with a lot of magic.
4. Went trough all your 500 posts, it was a funny reading (one could say “deranged psychology specimen frustrated with environment”). But no single post with backing facts. A lot of of insulting posts though (well, if I missed some post where you actualy said somthing intuitive I’m prepared to eat my words on this)
5. Even your nick is biased, insulting and childish (but it tells a lot about your psyche)
You remind me of people I tend to avoid as much as possible in my life. But when I start conversation with them, I know where conversation is heading
p.s. I tend to enjoy conversations like this (feeding the trolls with their own medicine, sadly for you I speak from experience not just promoting FUD)
p.p.s. You don’t remind me of anyone, because I can’t find as bad (maybe a better description is frustrated and childish) poster on OSNEWS as you are, well… maybe oGalaxyo
p.p.p.s. Either you won’t give answer or answer will be just as unintuitive as this one I’m responding to. In both cases I don’t give a f–k:)
Just test me.
I like how you instantly assume that I don’t know jack all about *nix. Sooner or later, in the course of a conversation with a Linux zealot, that always comes out. It is inevitable: “You know nothing about Linux, you are an idiot.”
Point proven. You just fell into your own trap. But yeah, go ahead, if you’re so convinced I know nothing, just ask. You might be a little surprised.
Ok, your named facts convinced me
Or not???? 
p.s. Another non-intuitive answer
Point proven. Thanks for playing.
Point proven. Thanks for playing.
Should I enclose my previous post in [sarcasm][/sarcasm] for you to understand?
OpenOffice’s codebase grew up as a proprietary one. Much of the code is over 12 years old.
If you want to make a case against ‘many eyes make bugs shallow,” you have to use a piece of software that was written from the ground up in an open fashion.
I would even discount software written by small organizations internally but released as open source.
I would instead look at a piece of software with an open development mailinglist/irc room. One that actively accepts patches from a good portion of its userbase.
Edited 2005-12-08 20:42
Agreed. For that matter an office suite is made up of a lot of code and not one is bug free. OpenOffice.org has it’s bugs, as does MS Office and every other office suite out there I can think of.
One of the reasons I use OpenOffice.org and StarOffice is because I encountered bugs in MS Office that are more annoying to me than the ones I encountered in OpenOffice.org and StarOffice so far.
OpenOffice’s codebase grew up as a proprietary one. Much of the code is over 12 years old.
It’s been open source since 2000. You’d think they’d gone over all of the code now.
If you want to make a case against ‘many eyes make bugs shallow,” you have to use a piece of software that was written from the ground up in an open fashion.
Apache. See Secunia advisories for Apache & IIS before replying.
“It’s been open source since 2000. You’d think they’d gone over all of the code now.”
Wow talk about ridiculously high expectations
“Apache. See Secunia advisories for Apache & IIS before replying.”
You expect apache to be perfect?
Once again………….
talk about ridiculously high expectations
The number of advisories per month on a software package may not be the most representative of it’s “goodness”.
I believe a more interesting comparison, may be the stack (IE: Linux, Apache, Oracle, Tomcat vs Windows, IIS, MS SQL, .NET) and man-hours of matenance per page served (probably thousandths would be an applicable scale). Couple that with compromises per page served and you would probably have a pretty clear picture of the situation.
I would even discount software written by small organizations internally but released as open source.
for example… OpenOffice, Mozilla products (based in Netscape), KDE (or at least QT), etc., right?
Unfortunately I think the Mozilla products suffer in the same way, though not on the same grand scale as OO.org.
ESR isn’t an idiot, he knows that the “Many eyes make bugs shallow” doesn’t always apply.
Let’s face it, there’s another culprit at work here: Large projects and organizations also build large bureaucracies, and it generally doesn’t matter if it’s for-profit or not. Just look at the UN, the EU or the US government, heck even Microsoft, they’re absolute giants in every sense of the word and yet none of them seem to do anything constructive or at least quickly and efficiently.
The bureaucratic effect of large projects simply cancels out the “many eyes” theory.
Read the airplane rule (ironic that it’s on esr’s site!): http://catb.org/~esr/jargon/html/A/airplane-rule.html
[quote]The bureaucratic effect of large projects simply cancels out the “many eyes” theory.[/quote]
I’m not sure that makes sense as a knock on OSS any more than as a knock against a large proprietary company. There’s probably less bureaucracy of this type with an OSS project, cuz turf protection is a leading cause of bureaucratic slow-down and is more readily found in a large proprietary company than in an OSS dev group.
If many eyes make bugs shallow, but OpenOffice.org 2.0 has a lot of bugs, then another suitable conclusion is that there are not enough eyes looking at the source.
If many eyes make bugs shallow, but OpenOffice.org 2.0 has a lot of bugs, then another suitable conclusion is that there are not enough eyes looking at the source.
Could be, but it is wrong in this case at least if you take your viewpoint on OSS community. Here is why
OO.o (July 2000) was derived from proprietary Sun StarOffice, which was derived from StarOffice by StarDivision (long back in 1986).
Meaning 14 years of not “many eyes” looking and 5 with “many eyes”. Writing your own code is much simpler than debugging code that someone else wrote. And difficulty rises with the size of the project.
Simple fact, OO.o has 10mio lines of source. Now the Grand Prize Question for you.
If you would aquire “many eyes”:) How many coders do you need to get over 10mio lines in 5 years at usual OSS approach, just to get once over the code? And how many to optimize 10mio lines of code?
OO.o is just a little to big (by size) mess, that was thrown into OSS to be able to be perfected in 5 years. Personally, I think better approach in that time would be complete rewrite by using SO code for template.
Source code makes it easier for programmers to fix other people’s softwares and contribute back bug reports and patches.
Two open source softwares that I currently use have been patched by me and I contributed back with bug reports and patches. It’s not always easy to get the patch in the repository, because I can fix one bug, but sometimes the architecure is what needs fixing, so the developers may hold the patches, bug it’s good to get the message that the software is buggy so the developers at least think about it.
I wonder how many programmers do the same? I would assume that most do it.
On Openoffice, it’s understandable that big projects need more people working on them. The problem is that big projects work counter-intuitive from the point of view of a programmer, who want’s a stable code base to work on, but big projects are sponsored by big companies who need features X, Y and Z for yesterday.
>> wonder how many programmers do the same? I would assume that most do it. <<
If I run into a bug, I try to work around it. More often, I’ll end up using proprietary software.
Yes, I’m a software engineer and I work for a very large tech company. But I don’t have any spare time to download OO, set up to work on it, fix it, debug it, test it, and merge my changes back.
How much spare time to I have? Not one minute. I am barely keeping my marriage together now. Once you leave school, get a real job and start a family, your spare time comes to an end.
I agree with the author. The FOSS conscept doesn’t scale very well.
I use OpenOffice at home.
I tried to write some macros, create some worksheet and databases, and I found numerous bugs that made me think OpenOfice was not seriously usable (v 1.2/1.3 a the time).
Next, I had to do similar stuff at work with Ms Office.
I encountered the same kind of bugs with office 2000.
I keep having trouble with MsOffice (disappearing files when opened from the network by to people, “hidden” functionality that are described in the help and that you found by clicking on a plane surface …)
I really think using complicated office suites is a real pain.
Um, there was no OpenOffice.org 1.2 or 1.3.
I think the post two levels up is referring to versions 1.1.2 and 1.1.3.
I can confirm that writing complex macros with OO’s spreadsheet component is impossibly useless.
I had one macro that would iterate over n rows, copy a result, paste it to another place; it always failed on the paste part. I do not know why, but OO complained about the lack of some Java dependency. This is especially interesting, because I tried it on Fedora Core 4 with GNU GCJ/Classpath, and it did not work either. I tried the same on Ubuntu using the same things, and no luck. I even installed Sun’s Java in both environments, instructed OO 2.0 to use it, and still no positive results.
I am thoroughly disappointed.
Call me whatever you will for wanting to iterate every 10 rows in a document consisting of 1,000+ rows, but it was absolutely necessary to finish some detail-heavy work.
“I really think using complicated office suites is a real pain.”
And along comes Apple with Pages…
Food for thought: the KISS rule is always right.
In the case of OpenOffice, it’s a big project with a core set of dedicated coders, hired by a company. I imagine most other FOSS coders wouldn’t bother hacking on it because
1. it’s so big that it may be hard to just jump into
2. there are already people being paid to work on it so why should someone spend their free/unpaid time on it?
Smaller (but useful) software that doesn’t have the backing of some foundation may be more likely to get all kinds of volunteer code work done on it. So bugs get found faster because
1. They are easier to track down
2. An experienced coder will realize that they are just as qualified to fix it as anyone else. They also may conclude that if they don’t fix it, no one else will.
And just to be a jerk, I have to say I don’t like Openoffice either. Abiword, Gnumeric, and hopefully soon, Criawips will be a complete replacement for me.
You need to understand what it is that you’re looking at before you can make a difference. I’d be willing to bet that most users of open source software do not have the skills required to be of any benefit in this regard.
What open office needs is to be divided and conquered:
Separate each program from the “suite” and work on them individually. Use the “suite program” as a program launcher. With open file formats implementing an Open Source OLE should be easier so you’ll have APPARENT integration to the end user.
With separate code bases for each program, those involved will have far less testing to do when a major change is made.
PS, the new Impress for Open Office 2 is awesome. I use it for every presentation I have at work and for flow charts for high level designs. It used to suck, especially around openoffice 1.1.4.
After hearing about how badass OO.o 2.0 was supposed to be, I installed it and took it for a spin. Specifically, I wanted to see if the database app might be a suitable replacement for MS Access. What I found was that in less than 5 minutes, I was able to lock up the program. When I actually got to the point where I could create a table, I got a message saying that Java was required. This seemed strange to me, considering that I have Java installed and working in all web browsers. Conclusion? I am not impressed.
Even Firefox, which is one of the better OSS desktop apps I’ve ever seen, isn’t exactly stellar (though I have not yet tried v1.5.) When you download and install an extension off the Mozilla update site and it breaks all menus and bookmarks, you gotta know something is not right.
It seems to me that OSS makes much better server software and desktop enviroments (KDE, Gnome, etc) then end user apps. Though there are a few good ones (K3b, Pan, Firefox, Thunderbird, etc), I just have to pass on most of them.
Even Firefox, which is one of the better OSS desktop apps I’ve ever seen, isn’t exactly stellar (though I have not yet tried v1.5.) When you download and install an extension off the Mozilla update site and it breaks all menus and bookmarks, you gotta know something is not right.
That’s not exactly Firefox’s problem, it has more to do with the extension, but I get your point.
It seems to me that OSS makes much better server software and desktop enviroments (KDE, Gnome, etc) then end user apps. Though there are a few good ones (K3b, Pan, Firefox, Thunderbird, etc), I just have to pass on most of them.
No, the problem with open source developers (generalizing here) is that they like more and more features rather than cleaning up after their mess. Where there’s 100 people building a program, you know you’re gonna have great (and not so great) ideas to implement. Though sometimes you have to draw a line and just work it out, instead of just feature packing it until it chocks. Everyone wants to see their code in a big project, but no one is willing to clean it up if it turns zup to break things.
I think Linus saw this coming so that now the kernel is freezed after 2 weeks(?) of feature packing. That’s how most software development works.
I think OO.o needs to remove some of its legacy code and modularize its code. Right now, it is almost impossible for anyone outside Sun to hack it due to its immense code size and compile time. It also has its own methods for painting and other basic tasks instead of relying on other libraries, which also needs constant upkeep.
As for Firefox, you should know that extensions are not part of Firefox. They are developed by *third parties*. If google toolbar consistently crashes IE, would you blame MS for this? The code in most extensions do not go through the same rigorous code review as Firefox and many will break Firefox. Note that downloading from MU does not mean the extension generally works. MU relies on volunteers to review the extensions and they are facing shortage of hands.
On a side-note, you should definitely try Firefox 1.5. It is much faster than its predecessor (try the back/forward buttons) and has preliminary SVG support (although somewhat incomplete). Many bugs are also fixed in this version. The best part, however, is the new automatic update system, which will automatically download small patches in the background.
Note that Gecko (the rendering engine used by Firefox and the Mozilla Suites) is basically written from scratch after Netscape donated their code. Right now, there are people actively hacking it outside the former employees of Netscape (esp. since javascript can be used to implement its functions)). This is in stark contrast with OO.o, which received very little code contribution from amateur programmers.
An example of a widely used Open Source program that greatly benefits from user feedback would have to be WINE.
Cedega, though commercial, goes even further with paying customers voting on what aspect of the program to improve.
I honestly wish ALL commercial software had a business model like that (asuming it was successful and could keep them in business.)
“Cedega, though commercial, goes even further with paying customers voting on what aspect of the program to improve.”
Voting is over-rated. I recall trying out Debian Hurd Cds on numerous occasions, and the packages that were included were chosen based on votes. Three CDs of mostly crap (many of which were X programs), with X11 on (IIRC) the last CD.
And let us not forget about Mr. Bush. Think conspiracy all you want, but I really believe that most people really are stupid enough to vote for the guy, not only once, but twice…
“And let us not forget about Mr. Bush. Think conspiracy all you want, but I really believe that most people really are stupid enough to vote for the guy, not only once, but twice…”
Stuipidity is not an exclusive trait of the conservatives. Many people vote down party lines without even knowing who the candidates are or what they stand for. Next time, try not to show so much bias.
Almost no one except a few Sun employees looks at the OO.org source base. Guess why there’s a big SUN logo on the OO.org startup screen; this is not a community project at all. The OO.org source is massive, incredible complex and almost impossible to build. It’s also a rather boring business app. Only things like the network stack which are interesting for hackers actually attract the “many eyes”. OO.org is an open source project only by the license, the whole “culture” around it doesn’t exist. It’s a really Sun product.
then why do all the OSS zealots view it as their “savior” in the office suite department? You say one thing, but the actions of the community say something completely different.
then why do all the OSS zealots view it as their “savior” in the office suite department? You say one thing, but the actions of the community say something completely different.
Firstly, I personally object to you categorizing a large lot of people into the negative category of “zealot”, but I will answer your question as to how I see it.
It is perhaps the closest thing to a replacement to MS Office. Similar enough to make Microsoft users not feel alienated. It is even packaged as a “suite”, just like MSO is.
So, I guess that is why the “OSS zealots” tend to respond with “OpenOffice.org” when asked for a replacement to Microsoft Office.
(edited for formatting – and again for ugly typo, oops!)
Edited 2005-12-08 21:22
Well, let’s be honest. OpenOffice isn’t the FOSS community’s savior, it’s the stopgap. The existence of OOo has made it possible for modestly technical users to use a free software platform as their primary desktop, and in the process enabled said platform to become good enough for many of the not-at-all technical users.
What I’m saying is, without Mozilla and OOo, the free software platform would have never achieved the basic level of desktop usability necessary to generate the hype that it has. It seems fair to suggest that some of the GNOME Office and KOffice developers wouldn’t be developing alternative FOSS office suites today if not for OOo.
Is OOo a long-term solution to office productivity on free software platforms? Maybe, but it seems more and more likely that the projects that grew up (predominantly) as open source offer a better evolutionary model for both users and developers. There is a benefit of tight integration between the office suite and the desktop environment that OOo cannot adequately provide due to cross-platform obligations. Their is a way to architect software such that community development models flourish. Linus, for example, demonstrated that highly modular, hierachically organized codebases can allow for distributed development without the devastating effects of Brooks’ Law. OOo is an example of Brook’s Law in all of its glory.
If KOffice had OOo’s MS Office filters today, I would immediately switch from GNOME to KDE. This is shaping up to be an outstanding office suite with great performance and a solid UI. There’s no reason to believe that OOo represents the FOSS community’s only shot at widely deployable office solution. Geeks have no subconscious loyalty, we will jump ship to another project just as soon as it makes sense. And everyone else will follow.
Edited 2005-12-08 21:54
OO is one of the reasons both KOffice and Gnome Office have stagnated. Both projects were started before StarOffice was opened up, and both projects were moving along at a good pace. OO removed much of the urgency, and it has only been in the past year that KOffice has taken off again. One reason for that has been undoubtedly the realization that OO is big time sink for anyone involved and something any hacker with good taste won’t want to touch for free.
Boudewijn Rempt
then why do all the OSS zealots view it as their “savior” in the office suite department? You say one thing, but the actions of the community say something completely different.
All? All? All? All what? Ah, yes, All OSS zealots.
Here you employ two fallacious bits of reasoning. First, that all of any group can somehow magically be construed as though they were a single person, and that anything said by one or even a few can therefore be regarded as having been said by all. This allows you to mistakenly characterize the entire group in whatever way you choose, good or bad, deserved or not.
Secondly and similarly, you characterize anyone who has the temerity to disagree with your personal views as flawed or defective. This allows you to mistakenly apply a label with a negative connotation so that everyone will understand that this entire group is to be regarded with contempt.
I could characterize this conduct as stupid, but more than that, it is so common. Lots of people do the same thing every day, never once realizing that there’s nothing new in the approach and the only ones impressed are those who already agree with you.
there’s a gigantic difference when you require people to PAY for the right to vote than letting anyone vote.
I know I’d listen to the people who keep the lights on and put food on the family (hahahah) over people who just have an opinion.
Lets face it folks, OO.o isn’t the only system with this problem. And more eyes and more money and more proprietary R&D and close knit developers aren’t the answer either.
Here’s the problem. Software has become so mind-boggingly complex that maybe there needs to be a grass-roots level to rethink the whole architecture of how it comes together. I believe that we are at a cross-roads in the regard in so far as the history and future of computing.
As a hobbyist programmer and professional chemical engineer, I find it interesting and confusing at the same time how professional developers argue about what language is best, e.g., Boo, VB, C#, C, C+, etc., etc. And inherent in this argument is the the technique employed, procedural, Object Oriented or the [new] Aspect Oriented approach. Then there is the argument over .NET versus Java thing. More of the same short term. All of these are long-term pain.
What we lack and what we need is a fresh modular approach to software architecture that enables the development of tomorrow’s massively complex systems using existing knowledgbase. We are struggling now with seemingly stupid bugs in packages that are essentially 15 years old, not because of a lack of intellect or design pattern know-how or tools, etc. but because the damn thing is a deck of cards assembled ina convoluted manner that defies anyone to touch it. To wit, amateurs like me run screeming on viewing the OO.o source. I can’t help with that. My God man, my remaining hair would vacate in an isntant! That’s not good, because I am one of those geeks with enough inspriation and [hopefully] talent to make a difference “on the side”. But, on the side aint good enough for OO.o. So what we have here is admisstion that the plan is broken. Let’s stop and think.
More eyes and more money and more investment isn’t the answer, we need to think outside the proverbial box because things are going to get a lot worse before they get better.
Remember the terms “feature bloat” was the in phrase to refer (generally in a hostile manner) to the myriad of [implicitly] stupid, unnecessary features software packages were providing in order to keep up with the competition? Well, I say that feature bloat is contemptuous reference by developers who are unable to give users want they want because the current development systems are too stressed to provide what we need.
We can’t continue for the next 15 years with the existing methods. Something as radical as the dawn of OO techniques ++ must be developed at a fundamental, architectural level. We need true modularity on a macro level, not at a mirco level provided through dynamic libraries and CLR-based systems.
I have chemical plants to design, build and run. I do this in my work. Our unit operations are all generlaly simple in and of themselves (I think so anyway). However, they collectively produce massively capable processes that yield results that are, at times, almost miraculous. We need to find a similar bridge to apply this to massively complex software systems like OO.o and Word, et. al.
New ClassbyRGCook : Inherits PreviousWork
Aint gonna cut it for the next generation.
Another rant concludes!
Good stuff here.
Although I generally agree with you, there are some “minor” problems.
_I have chemical plants to design, build and run. I do this in my work. Our unit operations are all generlaly simple in and of themselves (I think so anyway). However, they collectively produce massively capable processes that yield results that are, at times, almost miraculous. We need to find a similar bridge to apply this to massively complex software systems like OO.o and Word, et. al._
This above can illustrate root of many software design problems.
Chemical plant is in principle very easy to design – all chemical processes (like theoretical OOP objects) have clearly defined inputs and outputs (and events etc); they can be perfectly described as simple state machines (actually I’ve coded controllers for chemical processes many years ago). Similar approach is usable for various data management programs – which do not have interactive user interface (buttons for stop/start do not count).
It’s almost impossible design interactive software in similar manner – relations and possible interactions between software components AND user are too complex and dynamic. There are numerous ideas, theories, models, partial implementations and so on, which all are trying to solve this problem and advance programming to next, better level – no breakthrough so far. Programs become too complex and unmaintainable – or users are not satisfied with their functionality (or both
.
Nail hit on the head. Core level modularity and keeping design and implementation as simple as possible.
We can’t continue for the next 15 years with the existing methods. Something as radical as the dawn of OO techniques ++ must be developed at a fundamental, architectural level. We need true modularity on a macro level, not at a mirco level provided through dynamic libraries and CLR-based systems.
Well, part of the problem seems to be that the open-source community is very conservative and attached to C or C++.
They don’t even look at what’s around them in terms of languages. Recently there’s been a debate here over using Eiffel (SmartEiffel) and we could see that half the people don’t even care to learn what other possibilities exist. Miguel DeIcaza took a lot of heat because of Mono, too, by FUD spreaders that never understood it.
When you look at commercial software, like Rational Rose, you see that some of the problems you mention are really only a problem to the open-source community. Many in this community are UNIX die-hard system admins, which means they largely ignore some new software stuff. Others are so immersed in open-source only, that they never look to see what is being done in the other camp.
This community is too fixated in C and derived, and it’s just plain dumb and unproductive to develop software in such a glorified assembler. OTOH, when they finally get this point, they fall in the “java trap.”
We can’t continue for the next 15 years with the existing methods. Something as radical as the dawn of OO techniques ++ must be developed at a fundamental, architectural level. We need true modularity on a macro level, not at a mirco level provided through dynamic libraries and CLR-based systems.
Well, part of the problem seems to be that the open-source community is very conservative and attached to C or C++.
They don’t even look at what’s around them in terms of languages. Recently there’s been a debate here over using Eiffel (SmartEiffel) and we could see that half the people don’t even care to learn what other possibilities exist. Miguel DeIcaza took a lot of heat because of Mono, too, by FUD spreaders that never understood it.
When you look at commercial software, like Rational Rose, you see that some of the problems you mention are really only a problem to the open-source community. Many in this community are UNIX die-hard system admins, which means they largely ignore some new software stuff. Others are so immersed in open-source only, that they never look to see what is being done in the other camp.
This community is too fixated in C and derived, and it’s just plain dumb and unproductive to develop software in such a glorified assembler. OTOH, when they finally get this point, they fall in the “java trap.”
“Miguel DeIcaza took a lot of heat because of Mono, too, by FUD spreaders that never understood it.”
Some of us understand the situation just fine. We’re just waiting for the other shoe to drop so we all can do an “I told you so” in all your faces.*
*Childish I know, but then we’re talking about a group that labels people that don’t agree with them, “FUD spreaders” no matter how many times it’s been explained to them, so it all balances in the end.
That’s just one example why Open Source is misleading. It was never about “let people look at the source code to find bugs” or “let more people work on a program to create a better one”. That are just side effects who may appear in some cases but in other cases they won’t happen.
The goal was to give every single person the freedom in the digital world he is used to from the real world and the freedom everyone deserved in a free society.
“The goal was to give every single person the freedom in the digital world he is used to from the real world and the freedom everyone deserved in a free society.”
Well the problem with freedom is the other side of the coin. Responsability, and hence we have articles like we’re discussing.
ESR isn’t an idiot, he knows that the “Many eyes make bugs shallow” doesn’t always apply.
I think you are wrong, it always apply. But there are some limiting factors. The “many eyes” must have skill and understanding of the code, it’s not like the million monkeys and typewriter thing.
And the whole article are based on a flawed example. OO.org simply does not have “many eyes”. Easily seen with common sense and simple math, but then you don’t get a headline generating the same traffic. The amount of code per developer on OO.org are huge, or put the other way few eyes per line of code.
I’ve tried many times to use OO in a work environment but always end up imstalling MS Office on my linux machine with Crossover – version2 is no different. Despite all the (thoroughly misleading) reviews that OO reads MS Office format documents “almost flawlessly” this is just blatantly untrue – With v2 it’s actually rare that I open a .doc without it throwing in page breaks, muddling paragraphs and throwing out inserted graphics and diagrams – Powerpoint/Impress is the same story (only worse) any .ppt file with anything more complex than a few lines of text and again it gets garbaged.
A little less time spent on fancy functions which will hardly ever get used and a little more time spent on getting this ever so basic requirement sorted and I would be the first person to use OO 100%.
That “ever so basic requirement” is truly a Herculean task from a technical perspective. Reading MS Office file formats and rendering the document identically despite closed format specifications, proprietary font packages, and (presumably) massively differing internal data structures is nearly impossible.
So much work has gone into getting the MS Office filters as good as they are, that many other aspects of the suite have been sorely neglected. They have one again been improved for OOo-2.0, and I just don’t see how they could be made to perform any better.
I would say that “almost flawlessly” is a correct assessment. It opens MS Office 2003 documents about as well as MS Office 97 does, perhaps better. “Flawlessly” will never happen.
Despite all the (thoroughly misleading) reviews that OO reads MS Office format documents “almost flawlessly” this is just blatantly untrue
This is not my experience. I use OOo at home and have MS Office at work. I prepare documents at home and take them to work and vice-versa. I work on documents, spreadsheets and presentations. Fair enough I don’t use macros, work with only moderate sized spreadsheets, and believe on principle that presentations should look like the ones I prepared for 35mm slide shows a dozen years ago. Furthermore I don’t use the database functionality of either suite.
However I have found only minor formatting problems when transferring files between the two suites. I have also found that our spreadsheet templates developed exclusively on Excel (not by me) have worked flawlessly in Calc, including the the graphical plots (BTW they have also worked similarily in Gnumeric).
Finally my presentations have been well received, but maybe thats because they concentrate on scientific content rather than special effects and silly irrelevant animations.
Just take a look at Windows 95, 98, or Me to see examples of successful buggy software. MS won a monopoly with buggy software. MS may not have been able to maintain their monopoly if they hadn’t shifted over to the NT kernel, but Win9x is the poster child of buggy software and MS made a killing with it.
OOo has been successful because it has most of the features that users need and it runs natively on both Windows and Linux. It also opens most Office documents the way they were originally created. The trump card is the price. Even though there are a lot of things that could be improved about OOo, it still gets the job done for a lot of people at a much lower price than Office (some people don’t consider piracy to be a legitimate way of obtaining software).
For those considering OOo Base as an Access replacement, don’t bother unless you like alpha software. It works for some things, but it isn’t nearly as mature as Calc or Impress, much less Writer.
I’d be willing to bet that 95% of users of software don’t know how to write a proper bug report on any piece of software.
What’s your point?
Well that’s what tools like bug-buddy are for. I don’t think that OSS can be picky about the kinds of users they get. “World Domination” is like that.
More importantly is how much of a pain in the ass some bug reporting systems are.
First you need an account, then you have to make sure it’s not a duplicate bug report by searching through some sadistic search page, then if it’s new you have to go through another slightly less sadistic page to submit the bug report and wait for anywhere between hours and days for someone to reply, usually telling you you’re on a waiting list.
Don’t get me wrong, I love OSS, but I’ve had to go through bug submission web pages before and it was so miserable I’ve learned just to leave it to everyone else. When I can submit bug reports by e-mail again I’ll be a very happy guy.
BTW. I write some of my own programs as part of learning how to program, and I usually release them under the GPL or under a closed source freeware license. Either way I give out an e-mail address to send the bug reports to. Setting up a miserably poor designed, criptic torturous page for bug reports is a very good way not to get them.
It is unrealistic to expect bug reports to come through email for a large project. It would take forever to read through everything and the majority would be bug duplicates. You also wouldn’t have the info you need to test the bug since most users don’t understand scientific methodology.
I sadly have to agree with you. I write some extensions for firefox and I have accepted the fact that even though I have instructions all over my site explaining how to fill in a bug report I always get the same thing, “I have a problem with your extension it doesnt’ work on my pc.” I end up asking the same questions: What os, what browser version, what extension version, what were you doing when it happened, can you provide steps to repeat the bug? Usually people only answer one or two questions before sending the bug report. It is almost as if people want to help out but once they find out it will take some work they disappear.
The author admits around the end of the article that it doesn’t apply (not in so many words) since you have to sign papers to make any changes to OOo….
I don’t know why he even publishes the story when he knows full well that OOo has requirements which scare off probably half or more of the people who’d think of working on it.
Oh, I know why, so OSNews and other sites will repost his work and his paper will receive ad revenues.
I think the issue is that OpenOffice is very prominent, but very flawed. The fact that a not-so-great product is the most mature open-source office app reflects badly on Open Source, and makes it hard to defend. IMHO the best solution is to fix OO, or make something better. Easier said than done, of course…
There’s nothing indefensible about this, try reading the other replies!
OpenOffice.org had it’s origins as a closed source application, most of the code making it up is still from the old days when there was nothing bug StarOffice which by the way was for quite some time a proprietary office suite. It’s also impossible to have that much code without bugs, no office suite whether commercial or open source has absolutely zero bugs.
Finally people seem to be taking full advantage of this as a chance to stab at OSS for all it’s worth. Eric’s quote said that it made bugs shallow, not blatantly obvious.
Back to the drawing board for the lot of you that think this article along with your comments is going to kill off FOSS.
I don’t want to kill off FOSS, I just wish it had a better office app.
When I say “indefensible,” I mean that the office-apps category isn’t a great place to make a defensive stand for the viability of Open Source. As long as there’s nothing better, people will point to OO and say “Is that the best that Open Source can do?” And the answer right now is essentially “yes… for office apps.”
That part about killing FOSS wasn’t really directed at you, the topic seemed ideal for squeezing that in since others were trying to use this as munition against open source.
I don’t think “indefensible” was the word you were looking for though, with what I thought it meant I was quite confused.
Because no one is fired.
Oh dear, how naive. When was anyone in your company fired for shoddy code, poor service, or losing money? The corporate world does not work that way. The fellow sweeping the floor at the end of the day may well lose his job if he fails to clean a corner of the executive suite, but the CEO who lost his company many millions will assuredly still get his bonus at the end of the year, in the name of “retaining executive talent” and “remaining competitive.” Quality is derived from the few individuals who actually care and are competent. This is true of any organization. The higher along the chain of command you are the more likely that someone else will suffer for your bad decisions. And those at the top did not get generally get there by being the best and brightest engineers or having scientific minds. They have a far more prosaic quality: shameless sycophancy!
“The fellow sweeping the floor at the end of the day may well lose his job if he fails to clean a corner of the executive suite, but the CEO who lost his company many millions will assuredly still get his bonus at the end of the year, in the name of “retaining executive talent” and “remaining competitive.””
Your homework for today is to Google for “Conseco” and when you’re done with that, then try “Stereotype”.
Ah, an insurance company, staffed by diligent honest men and women, and led by visionary executives. I retract my statement, HP’s Carly Fiorina and GE’s Welch notwithstanding.
In all fairness, Word v.X for mac doesn’t even flawlessly open all word documents. If Microsoft can’t manage to have two products that open the same documents *with* the specs, it’s hard to imagine reverse engineering succeeding at it. My solution was to translate everything to RTF which, while a Microsoft format as well, is publicly specified and well supported by a huge number of apps.
I agree completely with the premise of this article, though. Open source is good for some things, but a massive code base like Open Office (must be, I’ve not seen it) is not one of those things. I agree also with the reply that said the problem is that it’s a *suite* – unbundle it and fix the pieces. IMHO Microsoft should try this approach too. The ‘integration’ of the suite doesn’t seem to buy you anything.
-HH
Increasing the number of competent developers looking at source might reduce bugs, but almost all users of open source are not developers. Seeing virtue in their looking at source is as logical as me claiming virture at looking at the inside of my TV set. If you don’t understand what you’re looking at, you’re not going to help fix it.
As usual, ESR picked up on the article and wrote a rebuttal on his blog.
The linky: http://esr.ibiblio.org/?p=239
Enjoy.
I also enjoy getting bug reports e-mailed to me, but that wouldn’t work if I had 16,000,000 users.
There’s a limit to the close-knit support of the person who wrote it helping you with your problem, and it’s more than one person needing help within the same 2 weeks to 2 years (depending on the developer).
Presentations which focus on the material are always wonderful, unless the material is just that boring
.
Presentations which focus on the visual aid (keyword: aid) always suck.
My priorities for an office suite are:
1. Price
2. Stability
3. Features
3. Performance
In my experience, OpenOffice 2.0 has a superb price, marginal stability, acceptable features, and utterly dreadful performance.
In my experience, Microsoft Office 2000 (the version I am most familiar with) has an utterly dreadful pricetag, excellent stability, excellent features, and superb performance.
I therefore use and recommend OpenOffice 2.0 only because MS Office is too ridiculously expensive.
I would actually give the exact same comments for my overall experience with Linux on the desktop vs. Windows XP Pro.
“Excellent” for Word stability is way too much. Word 2000 is what I earn my bread with, and I’m less than impressed with its stability. Other than that, I agree with your assessment and would recommend Word 2000 over OO.o 2.0 for any advanced and even not-so-advanced work.
Let’s say that from what I gathered from earlier posts, I see OO.o as a bloated piece of software with numerous bugs and a possible dead-end future. If I were going to “assemble” a “suite” of similar applications (give or take a couple if constrained), what would regular OSnews posters suggest as the contents of my non-suited suite? Is KOffice good enough? Would AbiWord plus X and Y be better?
Security expert Gene Spafford has pointed it out that Open Source in itself doesn’t really get you anything in terms of security. It does make it /possible/ for people to view the source code, but he notes that what matters is having people with the right technical skills and tools reviewing the source, rather than many eyes.
And it makes sense – if you have a million 10 year olds looking at your open source software, or a highly trained security professional reviewing closed source software, which do you think is more likely to be secure?
The other thing is that people have to be actually interested in a project to review it’s source. With closed source software, you can pay someone with the right skills to do this (whether your business management actually stumps up with the cash is entirely another matter of course!!)
When you have a variety of patches being submitted from various sources who rarely actually test their patches well in combination with other patches, you get what you have now.
The real issue here is the complexity of the two systems. OOo is making the same mistake as Microsoft in copying the same ‘feature set’ that MS Office provides – a feature set that most users simply don’t need. I believe that simpler software will produce more reliable software.
OOo ought to look at alternatives such as Atlantis/Nova (http://www.atlantiswordprocessor.com/en/) or AbiWord for good quality ‘simple’ word processing software. Gnumetric and JPS Development’s GS-Calc (http://www.jps-development.com/) – which is shareware are similar options for a spreadsheet.
Follow the old addage – keep it simple stupid!
Regards,
Peter
It really depends what you are trying to do. Everyone seems to use WP for all writing projects. This is not sensible. I have just completed an 80 page paper, properly structured, with TOC and foonotes and the usual layered headings, subsections and so on.
Can you do this in OO? Yes, but you are enormously hampered by the way that it confuses layout and composition, and I found that Outlining defeated me, given the time available. Can you do it in Word? Yes, but the same thing applies, Word’s outlining capability is a nightmare. Not to mention the issues about the stability of master documents, which have received widespread comment.
The problem with modern WP packages is that they have some of the features required to layout a quite sophisticated book – pagination across different files, footnotes and margin notes, all that stuff. But they also lack some features, and they have others which positively get in the way.
There is probably no solution to this within Office suites as they are presently defined. The answer may be: use WP for notes, letters, memos, short, fairly unstructured stuff. If you want a real writing tool, either use an outliner, or use Lyx. Or Mellel (for the Mac only).
People are talking here about KWord. Its a very fine page layout program, and will get better. Its totally useless as a tool for writing large structured documents. Abiword is equally useless. On the other hand, KWord or Abi word will do fine for memos, and KWord is great for DIY newsletters, brochures etc.
My bottom line is: OO is not terribly good for some kinds of writing, neither is Word. I don’t know that either is much worse or better than the other. But, used for what WP is actually good for, they are probably about as good or as bad as each other – equally over featured!
If you want a real writing tool, either use an outliner, or use Lyx. Or Mellel (for the Mac only).
I would like to add to this list TexMacs:
http://www.texmacs.org/
A great writing tool especially for mathematicians and physicists
Holy cow that is an awesome tool. I cant even figure out how to use Math in OOo. Mathype was great for me instead. OOo needs to make their software very easy and intuitive and I dont think they can. That is probably why Open Source OSes are still playing catch up in terms of popularity. Everyone knows Linux is great but to get the most requires a steep learning curve and people dont want that. One of the many reasons why a lot of BIG open source projects dont make it mainstream IMHO.
“OOo is making the same mistake as Microsoft in copying the same ‘feature set’ that MS Office provides – a feature set that most users simply don’t need.”
I don’t see this as a mistake, some people actually need and use these features, especially business and government. It’s true it’s still a problem that a lot of people don’t even know about the more advanced features and couldn’t be bothered to use them: however, most people have simpler needs – writing basic letters, memos, etc… but you don’t speak for everyone saying these features aren’t needed!
For writing longer documents, Word/OOo Writer won’t do everything you need, but KOffice/AbiWord are pretty useless for anything beyond simple documents. I think MSOffice/OOo do ably fill the niche between “professional/industrial” software and mere “memo-writing” programs.
Anyone played with the beta? Does it look like it has addressed a lot of the shortcomings of the Office suites? I use OOo cause it is free. While yes it has horrible performance, with some tweaks and the quickstarter enabled it is instant startup on my computers. I however see sometimes when I open a large MS word file it is painfully slow to edit on even a very high end computer as mine is….because of that I am probably going to buy me a copy of Office 12 and never ever buy any Office suite again.
My chief complaint about OO Calc, from my extremely light usage, is how poor its graph-making is. AFAICT, for an X-Y scatter plot, you can’t even assign one set of data to X and another to Y. You can’t set a custom window (xmin, xmax, and so forth). You can’t have it list just numbers along the X-axis to save space, instead of “row 1,” “row 2,” etc. for each tick. This is functionality one might use in high school, and yet…
I have found recently StarOffice 4.0 on one of my CDs. I installed it (must make some links) and I was amazed. It worked as fast as Office 95 on XP on P4 3.06GHz (I have Celeron 2400). And has it much more functions? What happend with the code through all these years?
If programmers did thing the UNIX-way (little good programs piped) instead of using the Windows-way (mammoth programs doing things half good) it would be better to maintain each program and, therefor, less bloated (you only install what you use), faster, easier to “jump in” and aid developing, …
OpenOffice.org is not a real open source project yet.
It looks like they opens up the project more and more, but it will most likely take some time. OpenOffice.org is developed in a closed way in the past. That is easy to see if you look at the code. This makes it harder for new developers to help.
OpenOffice.org 2.0 helped a lot, but it is still much work left to make it more open to new developers.
The title states the rhetorical question “If This X is a Success, Why Is It So buggy?”.
An example from real life: Windows – the operating systems with more security holes that swiss cheese.
Success and perfection does not (or even often) go hand in hand. I have too often seen superior technology perish before flawed technology with superior marketing.
Sorry for this shameless promotion, take a look my answer here: http://charles-blog.libervis.com
OpenOffice.org is not as open as other projects. That’s why it does not have ‘enough eyes’.
See KOffice source code online: webcvs.kde.org
See OpenOffice.org cvs? bug database? nope, register to take a look at them.
I cannot even do my math homework for the day because the damn chart function in the spreadsheet application just produces the outline of a chart but renders nothing inside of it. It does not actually plot the lines. How disappointing, I cannot even do my own homework with this thing.
Why can this thing not make even a simple x-y scatterplot or line plot?
SUN developers control Open Office.
Sun released Open Office to poke a rival in the eye.
Star Office doesnt have the bugs.
Do you trust SUN one or TWO companys who support SCO.
Why is java required at all.
Sun is trying to win Market share. Nothing new or different just market share.
Do I like the idea of Open Office YES, But until Open Office is completly under GNU/FOSS, not necessarly GNU/linux control
expect bugs.
Sun’ stuff all of the fillowing
Suns Patented Java Desktop
Suns CDDL
Suns Java
Suns Star Office
Suns Open Office Oh yes controlled by Sun Microsystems employees
http://www.yellowtab.com/images/news/ooo/dependencies.jpg
http://www.yellowtab.com/images/news/ooo/dependencies.jpg
Good lord, is that for real? O.O How do other projects’ dependency charts look?
Well alright! Phew, glad we got that cleared up! I guess we’ll just have to call it quits and abandon the Open Source movement huh? Well that was fun for a while, but the guy’s right, so back to a closed-source for us!
Honestly, this is a 10 year old experiment; it is still developing (no pun intended)! It’s not perfect yet, but who would expect it to be? This system is going to mature, I can only imagine in the future when all big projects have closed source and open source components to them. Again, maturity only comes with one thing; time.
When you talk about software that involves specialized knowledge, like word processors, than you’re talking about a different entry level. Some softwares, like compilers, get a lot of attention from specialists because they’re such a fundamental piece of the puzzle. Other projects (say, a compiler in a non-Algol-derived language) attracts less attention.
Many hackers will get people to join in projects that involve infrastruture (mail, etc) or systems programming, because many people are knowledgeable in that domain.
To hack on a word suite, it’s not enought to know the language (C++, etc). You must know *what* to code.
The same applies to very specialized software, like image processing or multimedia. For instance, the BBC wants to develop an open-source codec for streaming media. The prerequisite is that you’re able to understand wavelets. To do that, you must be knowledgeable in mathematics at least at the level of an engineering career to even begin to understand the documentation on wavelets.
In OO.org’s case, there’s a huge code base to be learned. This is the part where documentation really is crucial. If you don’t have by the beginning of the project, then you better take remedial steps and contract a technical writer (GNOME did this recently). If you project grows to a big size, you can’t expect help from people if they don’t even have a roadmap.
The part that ESR gets wrong is where he creates this mistique that coding is easy. The free software community is ridden with sloppy code, made by people with very little discipline in programming, as well as the opposite end, where you have the very best experts in something (OpenBSD hackers are the guys who most understand C programming in the open-source arena).