Yeah, I'll agree that the GPL does make it difficult to get equitable development funding costs shared around. Obviously, no GPL'ed software is developed for free. It would be nice to be able to have some reasonable way of spreading costs for it while still preserving the guarantee of openness and non-monopolization, but it's very unclear how that could work.
ESR is correct that a lot of open source software is developed by programmers working for other sorts of industries who need a piece of software to get their job done. For software that needs development beyond that point, it might be interesting to have a 'commercial for a couple of years' clause, which would allow commercial development and sale in traditional non-sharing style for a limited time, wherupon the code would revert back to the public commons.
It's the idea that I can only work on it on their terms that I have a problem with.
So? That sounds like the capitalist ideal of private property, to me. Would you rather that the state forced people to allow you to use their work in anything you like?
If your attitude is, "I'll make use of what people are stupid enough to give away, but I am not giving away *my* hard work", then Congratulations! No one will force you to! Just don't expect us to facilitate you doing what you want to do with our software.
So why on earth is it that your photograph on your home page shows that you use so much GPL'ed software, then? If you feel moral compunctions about free software sharing, then step up to the plate and stop using it.
Nonsense. The GPL doesn't deny anyone the ability to profit from the fruits of their labor, unless they choose to labor on GPL'ed software that they do not hold the copyright for. Believe me, no commercial software company will allow you to incorporate their code to create a derived work and then sell it, either. Not unless you pay them for the privilege, right?.
The presence of the GPL and code written under the GPL does not in any way prevent you from finding such a commercial vendor and offering them hundreds of thousands of dollars to allow you to base derived works on their products, if you like. The presence of the GPL and code written under the GPL does not prevent you from writing your own code from scratch, and profiting thereby.
I'm mystified at this kind of confusion that people have over the GPL. The people who complain about the GPL tend not to complain about the BSD license, when in fact they are complaining that they are not being allowed to profit adequately from someone else's work. Is that model more in keeping with your notion of capitalism? Take my work for free, and make money on it? If you are so concerned about the integrity of the capitalist system, pay cash on the barrel head for your software libraries, and be done with it.
Complaining that you're having to pay more than people who are willing to abide by the GPL is laughable for someone so concerned about Communism.
I think smart kids will handle the transitions
to more difficult languages okay. Programming in BASIC involved 'taking too many things for granted', perhaps, but every new language you learn should have something new to teach you anyway, or else why bother?
There are already tons of implementations of the JVM and the Java compiler, in case you hadn't noticed. Sun has a very very extensive language specification published, and anyone is free to
implement it, no strings attached.
The trick is that to make Java useful for most things, you need the very extensive libraries that are copyrighted by Sun.
But this is just the same thing as C#, isn't it? C# is made to bind tightly to the Win32 API set and to.NET. Having the C# language standardized will mean relatively little if all of the API's that C# code depends on are not standardized. And Microsoft has *always* laughed at anyone who suggested that they should 'standardize' their API's.
This must be the one of the few slashdot stories where the author of the story has more comments than the goat.se.cx trollers.;-)
As far as CORBA, Ganymede depends on the object graph serialization and distributed garbage collection features of RMI to provide a really tight binding between the GUI client and the server. Plus, RMI is simpler to work with, the
implementation that comes with the JDK has been
stable for a long time, and there is pretty much
ubiquitous support for it on the Java platform, making it possible to run the Ganymede client
just about everywhere without a lot of hassles.
As far as support for Python/C/C++/Perl, etc., there is an XML client that allows for high volume transactional data loading, and I imagine evolving that XML interface towards something like SOAP would be a better investment of development resources at this stage of the game.
Ganymede comes out-of-the-box with support for using the Java plugin on Linux, Solaris, and Windows. You can plug a Java 1.3 JVM into Netscape 4.7 and IE and the Ganymede web launcher will do the magic to invoke the proper plug-in and/or ActiveX control. If you are using Mozilla on Linux or Windows with a recent JRE, you can just use the 'native' option in the Ganymede web launcher and it'll pick up the JVM itself.
- jon
Re:Webserver won't play with netscape on linux
on
Ganymede 1.0 Released
·
· Score: 2
What, our web server? I run it, and I do everything from Linux. Works fine in Mozilla and Netscape 4.7. It might not work with IE, but then, that's someone else's problem, isn't it?;-)
You're probably just seeing it saturated. I bumped the number of concurrent web clients it will serve, but the proud Slashdot legions can probably still overwhelm it if they care to.
Many thanks for the comments, Jeff. I agree with you about the difficulty of handling transactions. Ganymede and GASH before it are
both really obsessed with providing a lot of interactive hand holding for the user, and Ganymede is especially obsessed with providing
really fine grained concurrency, and transactions seemed like the way to handle that, especially since things like NIS and DNS tend to require complete rebuilds when you make updates.
Ganymede is completely and totally my "second system", to quote Frederick Brooks. It is bigger and more complicated than it needs to be because I chose to optimize for some peculiar things. Ganymede is great for what it is, but what it is is in many ways kind of an odd duck. Its good points come out of our experience with GASH, so I do know that it works really very well for managment of a single domain network environment.
When I started writing Ganymede, I had never heard of the word 'metadirectory'. I only applied it after the fact, to try and express the fact that I intended Ganymede to manage other directory services, and not to be the primary directory service consulted by everything.
Thanks very much for your insights, I would have loved to have kicked these ideas around with you about five years ago.;-)
Some vendor tar programs can't handle very long filenames. Try using gnu tar. If you still have problems, check to make sure you downloaded the whole thing successfully.
I chose the back-end approach that I did for several reasons. First, when I started designing Ganymede back at the tail end of 1995, there were no GPL'ed SQL servers that supported transactions and appropriate locking. Second, from the predecessor project I developed a very strong fear of having to have my code deal with external files that could be corrupted/edited by someone. While the backend in Ganymede may not be terribly scalable for enterprise needs, I do at least have the assurance that the data has no real chance of being changed behind my back, and I can focus on controlling the changes made rather than trying to
worry about trying to make sense out of a random mess. Third, I wanted to be able to provide a reasonable object-based schema editing facility that could be used during runtime, and figuring out how to migrate object relational schemas with the tools that existed at the start of the project seemed forbidding. Fourth, I wanted to make Ganymede portable, and very easy to install and maintain with the limited resources I had to build the thing.
Similar reasons hold for the non-encrypted, non-PKI transport layer.
All that said, no, obviously the current choice of back-end is not appropriate for "enterprise" use if enterprise use means more than 50,000 users, say. I do imagine there are a lot of sites out there that could use some assistance dealing with their 10,000 users, though, which is sort of where I am aiming at with 1.0. I'd love to work with interested developers to try hack SleepyCat's Berkeley DB into Ganymede using JNI for 1.1, say. There's only so much that I can do myself, though, as I've got a lot of non-related systems administration tasks to do at work this summer for the people I work for. First up this morning, of course, was to increase MaxClients on our Apache server.;-)
Where I work, we master all of our accounts for
UNIX and NT, along with all of our email routing, our NFS volume definitions, our automounter configuration, and our DNS, in Ganymede.
We synchronize passwords from Ganymede into NIS, Samba, and our Windows NT PDC. We also
configure tacacs and radius, and LDAP from
the same master database. Ganymede provides
a high quality GUI interface, and allows you
to designate privileges over the directory
database to as many classes of administrators
as you like. Several people can be simultaneously browsing and making changes to the database,
and when transactions are committed, a background thread updates all of the network services.
Ganymede is the closest thing that the open
source community has to NDS or Active Directory, in terms of being a complete management solution, even though it is not based on LDAP and it
does not scale anywhere near as well as
an Active Directory or NDS. If you are looking
to manage a single location, however, Ganymede
will do the job right.
Take a look at the web site in my.signature, below. Ganymede 1.0 is due out within the next week, along with a userKit that supports password synchronization for UNIX, Samba, and Windows NT, with password quality checking handled by Clyde
Hoover's excellent nPasswd
passwd validation suite.
I think of this as directly analagous to
the DNS system. Once upon a time, NSI owned
the identity information for all top-level
domains under.com,.net,.org,.edu, etc. They
were forced to share the registration privileges
over these domains with other registrars in
a competitive framework. Hailstorm type
services need EXACTLY the same approach, where a user's identity could be a token like <xpp:id ref="jonabbey@burrow.org" reg="soap://microsoft.com/user/registry"> to indicate
a user registered at microsoft, where another,
equally valid identity token could be <xpp:id ref="jonabbey@burrow.org" reg="soap://aol.com/registry">.
The question is, who is going to bell the cat,
and create the sort of ambitious web services that
Microsoft is proposing, except without the Microsoft lock-in? Where is AOL and Sun and IBM on this?
Not necessarily, the passport-type account information could be stored on the client in the same way the Microsoft passport authentication information would be. DNS isn't made useless because there exist multiple registrars. For something of the scope that Microsoft is proposing for Hailstorm, I'd like to see competition to provide those services.
I have some friends who were working at a company that did a Wallet implementation for this sort of thing. It's *not* like Microsoft is the only company that can do something like this.
Sure, they want their services to be available on as many platforms as possible. My question was whether it would be possible for others to provide services with identical XML/SOAP interfaces and let consumers choose to use another provider for Hailstorm-type services, rather than to use Microsoft's, so when a user installs XP, he gets asked whether he wants to use Passport-type services from Microsoft, from AOL, or from Amazon.
Which I rather doubt Microsoft is going to encourage / allow, for all their talk about open services and interoperable standards. I could well be wrong about this, which is why I asked whether anyone did feel hopeful. If Microsoft did have the courage to allow competitors an even footing for the new services, I'd be much, much more sanguine about Microsoft. It's just that they never have allowed an even footing for anything like that.
And so should anyone be, to read that set of articles from BusinessWeek. They paint a picture of a Microsoft without limits to its control over the industry, and without limits as to its profit making power, feeding into greater control, feeding into greater profits..
Of course, the article doesn't even whisper the word 'Linux' or 'OpenSource' or, heck, even 'Java' anywhere. The picture they paint of Microsoft run rampant across the industry would be a completely, perfectly accurate one were it NOT for Linux and the Open Source world in general. Throw in the fact that XP will have the yummy corporate 'rights management' stuff built in, and you've got our biggest nightmare, right?
If the article had talked about how Linux has blunted the Windows 2000 server initiative, or about how Apache still runs most web servers, or about how there are dozens of manufacturers selling Java platforms, this would have seemed a good bit less scary. Fortunately. In my view, this article paints a clear picture, that we have three choices. One, the government slaps Microsoft down in some fashion, to impede its monopoly creation and maintenance ability. Two, Microsoft gets ever more powerful and buys pretty much whatever it wants to. Or, three, that everyone else involved in the industry works together on common standards for fear of their lives. That means Linux, that means Java, that means Mozilla, that means Ogg Vorbis, that means XML, that means an open AIM, that means standardized commodity streaming MPEG2 and MPEG4. And all of that might not be enough to forestall Microsoft if they become or remain the only ones with the ability to monetize the net effectively.
When it comes to service provisioning, the openness of the underlying software doesn't matter so much. Like Tim O'Reilly says, it's the openness of the web services that will matter greatly in the next phase of the net. If Microsoft makes their XML/SOAP protocol based services open enough that a Novell or an AOL or
an IBM or an Amazon or a Walmart or a Palm can compete to
provide Passport-type services integrated with XP, then perhaps Microsoft won't be such a threat to competition. Anyone feel hopeful?
Lyrics.ch is my all time favorite example of
the sort of world the RIAA and associated friends
would like to see. Their use of (signed, and thus
dangerously unsafe) Java to prevent cut and paste,
printing, or even screen shots (no control over
the scroll rate for you!) is just insulting.
Trust us to not be downloading non-sandboxed Java code to reformat your hard disk,
and we'll trust you to see the precious and secret lyrics to 'Happy Birthday' on your computer
screen momentarily. And, while you're here, why not click on our banner ads?
Happy Birthday to you, Happy Birthday to you, Happy Birthday dear citizen, Happy Birthday to you.
Myself, I'm one of those "rich" people who pays my share of those income taxes and probably a goodly chunk of yours, too, and I'd like to point out that it isn't the government's money, it's mine, and I'd like to be permitted to keep a little more of it.
Okay, you keep a little more of your money, I'll keep the roads, public education, environmental protection, judicial system, police officers, air traffic controllers, and military. How much money do you think you can make without those services being there?
People who talk about the government's money vs. the people's money don't generally tend to talk about the people's debts, or the people's services. Think government needs to be scaled back? Fine, that's a discussion we can have. Let's not act as if taxation is a priori theft, because it isn't.
Bush's new money goes to get our millitary people off of food stamps... yes you read that right).
Many military people qualify for food stamps because
they are paid little by civilian standards. One of the reasons they are paid little by civilian standards is that their housing is completely taken care of, and the value of their housing is not taken into account when calculating eligibility for Aid to Families with Dependent Children assistance. That doesn't mean that we have a military depending on AFDC to make sure their kids get fed. It is in significant part an artifact of the way compensation is given to the enlisted ranks.
Being as we're all concerned about blind statistics, and all.
1. We are using evidence that fits the conclusions of our assumption to prove our assumption. Bzzt, circular logic, thanks for playing Boole's Buzzer today.
Huh?
2. The massive infusion of new phyla in the Cambrian explosion cannot be explained by the isolation of breeding stocks. Even though the planet was just recovering from a "Snowball Earth" episode, the are widespread biodeposits during this period, indicating a stable ecosystem filled with complex organisms. What's more, this stable ecosystem didn't have to endure Darwin's millions-of-years "cut and try" approach, it was stable early on in the 2-million-year window about 480 million years ago.
That doesn't in any way contradict the notion
that separation of breeding pools facilitates speciation. The Cambrian explosion is fascinating
and there is obviously a lot to be studied there, but who ever said that evolution was allowed only one trick?
It's manifestly obvious that the process of evolution is susceptible to contingency. It happens all the time.. a virus mutates in a way that allows it access to an entirely new population of carriers, or it is merely brought into close enough contact with a new and vulnerable population, and you have plague. That plague may wipe out 75% or more of a population, giving tremendous evolutionary advantage to those whose genetics happens to afford some resistance or resiliance to the infection.
It is clear that discrete events can occur that force the ecological system out of a state of stability and into a chaotic period where living organisms have to adapt to the new reality, or perish (rabbits in Australia, anyone?). That is not inconsistent with the idea that separation of breeding pools can facilitate speciation. Biology is one of the most complex of subjects, and the number of processes and mechanisms that swirl around in the ecology are legion. Every mutation has an opportunity to change the playing field to a greater or lesser extent.
But, wait, that's a contradictory statement! IF speciation is driven by a mindless process, it should be spontaneous, in the sense of occuring without outside intervention. Just species popping up all the time, wherever the process is allowed to work its special magic of DNA differentiation. Here's the problem: we have palentogical evidence of discrete speciation EVENTS, not a continuous ebb and flow of species.
Perhaps I misspoke. I meant 'spontaneous' to mean 'without cause' and some shading of 'instantaneous'. The notion of 'outside intervention' is so unnecessary to my understanding of the world that it never occurred to me that one would read 'spontaneous' as meaning 'without outside intervention'. I believe speciation is always for cause, and that if we are fortunate enough to have intimate knowledge of a particular population as it goes through a speciation event that the reasons for the speciation can generally be observed.
In addition, two points about speciation events. First, speciation events may be slow (or fast) on a human timescale, but they are bound to be very fast on the geological time scale that most paleontology deals with. Second, if you stretch your point of view out long enough, you'll see the continual process of the ebb and flow of species I have been speaking of. Not a constant ebb and flow, in the sense of an unvarying rate of change, but continual and continuous in the sense that the process of species being born and destroyed has never yet ceased on Earth.
Something is wrong with the model, and facing the facts compells us to admit it.
Well, something is wrong with the simplified model we are talking about here, yes. I'd recommend Darwin's Dangerous Idea by Daniel C. Dennett if you want a really good description of the structural nature of species and the process by which a genetic population drifts and clumps. It intelligently deals with some very important questions, like, 'what is a species?', that are really essential background for this kind of discussion.
No, sir. I think the Cambrian Explosion is played down in the textbooks, for this very reason: it stands ready to skewer some very sacred cows.
It is a rather unworthy argument to simply fall back on calling your opponents cowards and liars. If there is a good argument to be made against some particular facet of evolutionary theory, it would be better to spend your energy finding the evidence to demonstrate your point of view, rather than disparaging others. It's hard work, because there is a mountain of evidence on the side of the modern evolutionary synthesis, but if you find something that others have missed you will surely make a great contribution to the world of science. If it turns out that you learn enough to doubt your current position, then that would also be valuable. Sacred cows certainly are troublesome, no matter whose side they are on.
After all, that bit about a million billion trillion trillion experiments that went on before is assuming that each experiment takes a year to run. Some will take longer, most will take far less. Let's say that every thing living today benefits from a lineal subset of the million trillion trillion trillion experiments that went before.
Random chance throws some variation into those experiments, but I don't call the fact that a million trillion trillion trillion experiments has led to some good results chance, I call that inevitability.
After all, there are 6 billion humans alone, and every one of those humans harbors on average 10^14 bacteria, or ten times the number of actual
human cells. For human-hosted bacteria alone,
that makes 6*10^23 bacteria, or
600,000,000,000,000,000,000,000. That's
600 billion trillion bacteria living on humans
alone. If we count all organisms living on the
planet, that figure will probably go up by at least another factor of billion, or more.
So nature is running perhaps a million trillion trillion experiments at once, or more. There is no mystery at all that with that much going on at any given time, and with each of those experiments building on the successful results of some of the million billion trillion trillion experiments that went before, that nature should figure out how to let a bacteria fuck with a wasp's reproduction system.
Something like this came to evolve in the same way that everything else did. Nature runs *trillions* of experiments at a time, ruthlessly ditching the results of the ones which don't measure up, which is the vast majority of those trillions. A very small minority of those trillions get to reproduce, and their offspring get to participate in the nex round of the game.
Nature has been running this game for billions of years. Some of those experiments, especially for bacteria, can take as little as 20 minutes to run. That's 26,280 generations a year, times 3 billion year or so. *Lots* of time.
Mankind hasn't invested one quadrillionth of the effort or resources in developing living things that nature has, and I can't imagine we ever will.
Ask Paul Erlich, who said in his book Extinction: "We have yet to see a single instance of speciation in the animal kingdom".
Well, even assuming that was correct when he wrote that, that was twenty years ago. Maybe science
has spent that time discovering new things? The way science works, a twenty year old book never has any sort of veto authority over new facts and new discoveries.
And, in any case, nothing in this wasp/bacteria tale deals with 'spontaneous' speciation, it deals with an identifiable factor which is providing the
barrier to breeding that can facilitate a process of speciation. A process, not a miraculous, spontaneous event.
It just doesn't happen in animals these days. Happens in plants, sure, but they have a different DNA structure than us. Come on, give up on the pipe dream of spontaneous speciation as an explanation for the diversity of species. It's beginning to sound as dated as spontaneous generation, which Pasteur and others blew away 150 years ago.
Oh. Care to explain what features of the DNA structure of the animal kingdom it is that prevents new species from being created?
Because, from everything I've read, species creation has been a constant process over the last 3 or 4 billion years or so. Existing species die out, leaving ecological niches where a new species can get a toehold, and the process of mutation and selection gives rise to species that can take advantage of the opportunity.
Slashdot Mirror
Yeah, I'll agree that the GPL does make it difficult to get equitable development funding costs shared around. Obviously, no GPL'ed software is developed for free. It would be nice to be able to have some reasonable way of spreading costs for it while still preserving the guarantee of openness and non-monopolization, but it's very unclear how that could work.
ESR is correct that a lot of open source software is developed by programmers working for other sorts of industries who need a piece of software to get their job done. For software that needs development beyond that point, it might be interesting to have a 'commercial for a couple of years' clause, which would allow commercial development and sale in traditional non-sharing style for a limited time, wherupon the code would revert back to the public commons.
This wouldn't be any sort of GPL, of course.
- jon
So? That sounds like the capitalist ideal of private property, to me. Would you rather that the state forced people to allow you to use their work in anything you like?
If your attitude is, "I'll make use of what people are stupid enough to give away, but I am not giving away *my* hard work", then Congratulations! No one will force you to! Just don't expect us to facilitate you doing what you want to do with our software.
- jon
So why on earth is it that your photograph on your home page shows that you use so much GPL'ed software, then? If you feel moral compunctions about free software sharing, then step up to the plate and stop using it.
- jon
Nonsense. The GPL doesn't deny anyone the ability to profit from the fruits of their labor, unless they choose to labor on GPL'ed software that they do not hold the copyright for. Believe me, no commercial software company will allow you to incorporate their code to create a derived work and then sell it, either. Not unless you pay them for the privilege, right?.
The presence of the GPL and code written under the GPL does not in any way prevent you from finding such a commercial vendor and offering them hundreds of thousands of dollars to allow you to base derived works on their products, if you like. The presence of the GPL and code written under the GPL does not prevent you from writing your own code from scratch, and profiting thereby.
I'm mystified at this kind of confusion that people have over the GPL. The people who complain about the GPL tend not to complain about the BSD license, when in fact they are complaining that they are not being allowed to profit adequately from someone else's work. Is that model more in keeping with your notion of capitalism? Take my work for free, and make money on it? If you are so concerned about the integrity of the capitalist system, pay cash on the barrel head for your software libraries, and be done with it.
Complaining that you're having to pay more than people who are willing to abide by the GPL is laughable for someone so concerned about Communism.
- jon
I think smart kids will handle the transitions to more difficult languages okay. Programming in BASIC involved 'taking too many things for granted', perhaps, but every new language you learn should have something new to teach you anyway, or else why bother?
- jon
There are already tons of implementations of the JVM and the Java compiler, in case you hadn't noticed. Sun has a very very extensive language specification published, and anyone is free to implement it, no strings attached.
The trick is that to make Java useful for most things, you need the very extensive libraries that are copyrighted by Sun.
But this is just the same thing as C#, isn't it? C# is made to bind tightly to the Win32 API set and to .NET. Having the C# language standardized will mean relatively little if all of the API's that C# code depends on are not standardized. And Microsoft has *always* laughed at anyone who suggested that they should 'standardize' their API's.
- jon
This must be the one of the few slashdot stories where the author of the story has more comments than the goat.se.cx trollers. ;-)
As far as CORBA, Ganymede depends on the object graph serialization and distributed garbage collection features of RMI to provide a really tight binding between the GUI client and the server. Plus, RMI is simpler to work with, the implementation that comes with the JDK has been stable for a long time, and there is pretty much ubiquitous support for it on the Java platform, making it possible to run the Ganymede client just about everywhere without a lot of hassles.
As far as support for Python/C/C++/Perl, etc., there is an XML client that allows for high volume transactional data loading, and I imagine evolving that XML interface towards something like SOAP would be a better investment of development resources at this stage of the game.
- jon
Ganymede comes out-of-the-box with support for using the Java plugin on Linux, Solaris, and Windows. You can plug a Java 1.3 JVM into Netscape 4.7 and IE and the Ganymede web launcher will do the magic to invoke the proper plug-in and/or ActiveX control. If you are using Mozilla on Linux or Windows with a recent JRE, you can just use the 'native' option in the Ganymede web launcher and it'll pick up the JVM itself.
- jon
What, our web server? I run it, and I do everything from Linux. Works fine in Mozilla and Netscape 4.7. It might not work with IE, but then, that's someone else's problem, isn't it? ;-)
You're probably just seeing it saturated. I bumped the number of concurrent web clients it will serve, but the proud Slashdot legions can probably still overwhelm it if they care to.
- jon
Many thanks for the comments, Jeff. I agree with you about the difficulty of handling transactions. Ganymede and GASH before it are both really obsessed with providing a lot of interactive hand holding for the user, and Ganymede is especially obsessed with providing really fine grained concurrency, and transactions seemed like the way to handle that, especially since things like NIS and DNS tend to require complete rebuilds when you make updates.
Ganymede is completely and totally my "second system", to quote Frederick Brooks. It is bigger and more complicated than it needs to be because I chose to optimize for some peculiar things. Ganymede is great for what it is, but what it is is in many ways kind of an odd duck. Its good points come out of our experience with GASH, so I do know that it works really very well for managment of a single domain network environment.
When I started writing Ganymede, I had never heard of the word 'metadirectory'. I only applied it after the fact, to try and express the fact that I intended Ganymede to manage other directory services, and not to be the primary directory service consulted by everything.
Thanks very much for your insights, I would have loved to have kicked these ideas around with you about five years ago. ;-)
- jon
Some vendor tar programs can't handle very long filenames. Try using gnu tar. If you still have problems, check to make sure you downloaded the whole thing successfully.
- jon
All reasonable, excellent, and obvious points.
I chose the back-end approach that I did for several reasons. First, when I started designing Ganymede back at the tail end of 1995, there were no GPL'ed SQL servers that supported transactions and appropriate locking. Second, from the predecessor project I developed a very strong fear of having to have my code deal with external files that could be corrupted/edited by someone. While the backend in Ganymede may not be terribly scalable for enterprise needs, I do at least have the assurance that the data has no real chance of being changed behind my back, and I can focus on controlling the changes made rather than trying to worry about trying to make sense out of a random mess. Third, I wanted to be able to provide a reasonable object-based schema editing facility that could be used during runtime, and figuring out how to migrate object relational schemas with the tools that existed at the start of the project seemed forbidding. Fourth, I wanted to make Ganymede portable, and very easy to install and maintain with the limited resources I had to build the thing.
Similar reasons hold for the non-encrypted, non-PKI transport layer.
All that said, no, obviously the current choice of back-end is not appropriate for "enterprise" use if enterprise use means more than 50,000 users, say. I do imagine there are a lot of sites out there that could use some assistance dealing with their 10,000 users, though, which is sort of where I am aiming at with 1.0. I'd love to work with interested developers to try hack SleepyCat's Berkeley DB into Ganymede using JNI for 1.1, say. There's only so much that I can do myself, though, as I've got a lot of non-related systems administration tasks to do at work this summer for the people I work for. First up this morning, of course, was to increase MaxClients on our Apache server. ;-)
- jon
Where I work, we master all of our accounts for UNIX and NT, along with all of our email routing, our NFS volume definitions, our automounter configuration, and our DNS, in Ganymede.
We synchronize passwords from Ganymede into NIS, Samba, and our Windows NT PDC. We also configure tacacs and radius, and LDAP from the same master database. Ganymede provides a high quality GUI interface, and allows you to designate privileges over the directory database to as many classes of administrators as you like. Several people can be simultaneously browsing and making changes to the database, and when transactions are committed, a background thread updates all of the network services. Ganymede is the closest thing that the open source community has to NDS or Active Directory, in terms of being a complete management solution, even though it is not based on LDAP and it does not scale anywhere near as well as an Active Directory or NDS. If you are looking to manage a single location, however, Ganymede will do the job right.
Take a look at the web site in my .signature, below. Ganymede 1.0 is due out within the next week, along with a userKit that supports password synchronization for UNIX, Samba, and Windows NT, with password quality checking handled by Clyde
Hoover's excellent nPasswd
passwd validation suite.
- jon
I think of this as directly analagous to the DNS system. Once upon a time, NSI owned the identity information for all top-level domains under .com, .net, .org, .edu, etc. They
were forced to share the registration privileges
over these domains with other registrars in
a competitive framework. Hailstorm type
services need EXACTLY the same approach, where a user's identity could be a token like <xpp:id ref="jonabbey@burrow.org" reg="soap://microsoft.com/user/registry"> to indicate
a user registered at microsoft, where another,
equally valid identity token could be <xpp:id ref="jonabbey@burrow.org" reg="soap://aol.com/registry">.
The question is, who is going to bell the cat, and create the sort of ambitious web services that Microsoft is proposing, except without the Microsoft lock-in? Where is AOL and Sun and IBM on this?
- jon
Not necessarily, the passport-type account information could be stored on the client in the same way the Microsoft passport authentication information would be. DNS isn't made useless because there exist multiple registrars. For something of the scope that Microsoft is proposing for Hailstorm, I'd like to see competition to provide those services.
I have some friends who were working at a company that did a Wallet implementation for this sort of thing. It's *not* like Microsoft is the only company that can do something like this.
- jon
Sure, they want their services to be available on as many platforms as possible. My question was whether it would be possible for others to provide services with identical XML/SOAP interfaces and let consumers choose to use another provider for Hailstorm-type services, rather than to use Microsoft's, so when a user installs XP, he gets asked whether he wants to use Passport-type services from Microsoft, from AOL, or from Amazon.
Which I rather doubt Microsoft is going to encourage / allow, for all their talk about open services and interoperable standards. I could well be wrong about this, which is why I asked whether anyone did feel hopeful. If Microsoft did have the courage to allow competitors an even footing for the new services, I'd be much, much more sanguine about Microsoft. It's just that they never have allowed an even footing for anything like that.
- jon
And so should anyone be, to read that set of articles from BusinessWeek. They paint a picture of a Microsoft without limits to its control over the industry, and without limits as to its profit making power, feeding into greater control, feeding into greater profits..
Of course, the article doesn't even whisper the word 'Linux' or 'OpenSource' or, heck, even 'Java' anywhere. The picture they paint of Microsoft run rampant across the industry would be a completely, perfectly accurate one were it NOT for Linux and the Open Source world in general. Throw in the fact that XP will have the yummy corporate 'rights management' stuff built in, and you've got our biggest nightmare, right?
If the article had talked about how Linux has blunted the Windows 2000 server initiative, or about how Apache still runs most web servers, or about how there are dozens of manufacturers selling Java platforms, this would have seemed a good bit less scary. Fortunately. In my view, this article paints a clear picture, that we have three choices. One, the government slaps Microsoft down in some fashion, to impede its monopoly creation and maintenance ability. Two, Microsoft gets ever more powerful and buys pretty much whatever it wants to. Or, three, that everyone else involved in the industry works together on common standards for fear of their lives. That means Linux, that means Java, that means Mozilla, that means Ogg Vorbis, that means XML, that means an open AIM, that means standardized commodity streaming MPEG2 and MPEG4. And all of that might not be enough to forestall Microsoft if they become or remain the only ones with the ability to monetize the net effectively.
When it comes to service provisioning, the openness of the underlying software doesn't matter so much. Like Tim O'Reilly says, it's the openness of the web services that will matter greatly in the next phase of the net. If Microsoft makes their XML/SOAP protocol based services open enough that a Novell or an AOL or an IBM or an Amazon or a Walmart or a Palm can compete to provide Passport-type services integrated with XP, then perhaps Microsoft won't be such a threat to competition. Anyone feel hopeful?
- jon
Lyrics.ch is my all time favorite example of the sort of world the RIAA and associated friends would like to see. Their use of (signed, and thus dangerously unsafe) Java to prevent cut and paste, printing, or even screen shots (no control over the scroll rate for you!) is just insulting. Trust us to not be downloading non-sandboxed Java code to reformat your hard disk, and we'll trust you to see the precious and secret lyrics to 'Happy Birthday' on your computer screen momentarily. And, while you're here, why not click on our banner ads?
Happy Birthday to you, Happy Birthday to you, Happy Birthday dear citizen, Happy Birthday to you.
Okay, you keep a little more of your money, I'll keep the roads, public education, environmental protection, judicial system, police officers, air traffic controllers, and military. How much money do you think you can make without those services being there?
People who talk about the government's money vs. the people's money don't generally tend to talk about the people's debts, or the people's services. Think government needs to be scaled back? Fine, that's a discussion we can have. Let's not act as if taxation is a priori theft, because it isn't.
Many military people qualify for food stamps because they are paid little by civilian standards. One of the reasons they are paid little by civilian standards is that their housing is completely taken care of, and the value of their housing is not taken into account when calculating eligibility for Aid to Families with Dependent Children assistance. That doesn't mean that we have a military depending on AFDC to make sure their kids get fed. It is in significant part an artifact of the way compensation is given to the enlisted ranks.
Being as we're all concerned about blind statistics, and all.
Huh?
That doesn't in any way contradict the notion that separation of breeding pools facilitates speciation. The Cambrian explosion is fascinating and there is obviously a lot to be studied there, but who ever said that evolution was allowed only one trick?
It's manifestly obvious that the process of evolution is susceptible to contingency. It happens all the time.. a virus mutates in a way that allows it access to an entirely new population of carriers, or it is merely brought into close enough contact with a new and vulnerable population, and you have plague. That plague may wipe out 75% or more of a population, giving tremendous evolutionary advantage to those whose genetics happens to afford some resistance or resiliance to the infection.
It is clear that discrete events can occur that force the ecological system out of a state of stability and into a chaotic period where living organisms have to adapt to the new reality, or perish (rabbits in Australia, anyone?). That is not inconsistent with the idea that separation of breeding pools can facilitate speciation. Biology is one of the most complex of subjects, and the number of processes and mechanisms that swirl around in the ecology are legion. Every mutation has an opportunity to change the playing field to a greater or lesser extent.
Perhaps I misspoke. I meant 'spontaneous' to mean 'without cause' and some shading of 'instantaneous'. The notion of 'outside intervention' is so unnecessary to my understanding of the world that it never occurred to me that one would read 'spontaneous' as meaning 'without outside intervention'. I believe speciation is always for cause, and that if we are fortunate enough to have intimate knowledge of a particular population as it goes through a speciation event that the reasons for the speciation can generally be observed.
In addition, two points about speciation events. First, speciation events may be slow (or fast) on a human timescale, but they are bound to be very fast on the geological time scale that most paleontology deals with. Second, if you stretch your point of view out long enough, you'll see the continual process of the ebb and flow of species I have been speaking of. Not a constant ebb and flow, in the sense of an unvarying rate of change, but continual and continuous in the sense that the process of species being born and destroyed has never yet ceased on Earth.
Well, something is wrong with the simplified model we are talking about here, yes. I'd recommend Darwin's Dangerous Idea by Daniel C. Dennett if you want a really good description of the structural nature of species and the process by which a genetic population drifts and clumps. It intelligently deals with some very important questions, like, 'what is a species?', that are really essential background for this kind of discussion.
It is a rather unworthy argument to simply fall back on calling your opponents cowards and liars. If there is a good argument to be made against some particular facet of evolutionary theory, it would be better to spend your energy finding the evidence to demonstrate your point of view, rather than disparaging others. It's hard work, because there is a mountain of evidence on the side of the modern evolutionary synthesis, but if you find something that others have missed you will surely make a great contribution to the world of science. If it turns out that you learn enough to doubt your current position, then that would also be valuable. Sacred cows certainly are troublesome, no matter whose side they are on.
After all, that bit about a million billion trillion trillion experiments that went on before is assuming that each experiment takes a year to run. Some will take longer, most will take far less. Let's say that every thing living today benefits from a lineal subset of the million trillion trillion trillion experiments that went before.
Random chance throws some variation into those experiments, but I don't call the fact that a million trillion trillion trillion experiments has led to some good results chance, I call that inevitability.
After all, there are 6 billion humans alone, and every one of those humans harbors on average 10^14 bacteria, or ten times the number of actual human cells. For human-hosted bacteria alone, that makes 6*10^23 bacteria, or 600,000,000,000,000,000,000,000. That's 600 billion trillion bacteria living on humans alone. If we count all organisms living on the planet, that figure will probably go up by at least another factor of billion, or more.
So nature is running perhaps a million trillion trillion experiments at once, or more. There is no mystery at all that with that much going on at any given time, and with each of those experiments building on the successful results of some of the million billion trillion trillion experiments that went before, that nature should figure out how to let a bacteria fuck with a wasp's reproduction system.
Something like this came to evolve in the same way that everything else did. Nature runs *trillions* of experiments at a time, ruthlessly ditching the results of the ones which don't measure up, which is the vast majority of those trillions. A very small minority of those trillions get to reproduce, and their offspring get to participate in the nex round of the game.
Nature has been running this game for billions of years. Some of those experiments, especially for bacteria, can take as little as 20 minutes to run. That's 26,280 generations a year, times 3 billion year or so. *Lots* of time.
Mankind hasn't invested one quadrillionth of the effort or resources in developing living things that nature has, and I can't imagine we ever will.
Well, even assuming that was correct when he wrote that, that was twenty years ago. Maybe science has spent that time discovering new things? The way science works, a twenty year old book never has any sort of veto authority over new facts and new discoveries.
And, in any case, nothing in this wasp/bacteria tale deals with 'spontaneous' speciation, it deals with an identifiable factor which is providing the barrier to breeding that can facilitate a process of speciation. A process, not a miraculous, spontaneous event.
Oh. Care to explain what features of the DNA structure of the animal kingdom it is that prevents new species from being created?
Because, from everything I've read, species creation has been a constant process over the last 3 or 4 billion years or so. Existing species die out, leaving ecological niches where a new species can get a toehold, and the process of mutation and selection gives rise to species that can take advantage of the opportunity.
Nature hates a vacuum, as they say.