Anyway, I think Yaztromo just solved his problem #1 - by getting his question posted on Slashdot, I don't think he'll have any trouble marketing his project now, assuming it's any good.
Just FYI, now that the SourceForge statistics for July 14th (the date this story was posted to/.) are available, I decided to look them up. According to them, our homepage views were a grand total of 83 for the day.
Perhaps the results for the first full day the article was online (July 15th) will be different, but thus far having this question posted to Slashdot hasn't caused the "slashdot effect" to occur to our website at http://www.jsyncmanager.org.
I'm not disappointed by any means -- getting my question posted wasn't intended to be a marketing gimmick in and of itself. Maybe the results for yesterday, the 15th, will be more suprising (one way or another:) ).
a 5 line shell script will take care of getting any libraries they need from the web. There is no reason your users should have to surf to get the libs. Get them for them.
That works well for Linux users, but not so well for OS/2 or Windows users.
Part of the difficulty is also that some of these libraries have a native component, and need to be installed correctly. On Linux, some of them need to be installed as root. Additionally, at least one library (the Java Communications API for OS/2) requires a web password to download it.
Having a shell script that just downloads everything is a great idea if we were developing Linux-only software, but we're not. That shell script isn't going to run correctly outside the Unix arena, and the facilities you (and I) take for granted on Linux to make this sort of thing happen don't necessarily exist on these other platforms (and if they do, are another seperate download).
I do thank you for the otherwise good idea (I'm not trying to put it down!), but I don't see it working terribly well here for all our potential users.
If you run out moneys then you can go live in friendlier country based on socialism or communism instead of rude arrogant capitalist system which force people to work to get what they want.
Just ask any other Americans.
Are there any Americans in here who can tell me what this guy is talking about?
I don't know what he's thinking, but I do live in a friendlier country that is socially liberal and fiscally conservative -- I'm Canadian.
I hope he didn't just automatically assume I'm American just because I'm posting to US-centric./:).
Not to put words in the poster's mouth, but the reason many people find NFS to be somewhat complex is that the protocol requires user and group ids to be synchronized between the client and the server. If you're talking about two Unix systems, this generally means running NIS, NIS+, or LDAP, each of which is more complex than a standalone/etc/passwd file.
Nonsense I say!:). You don't need to run any extra special services like this for a small home installation -- just make sure that if a user exists on all your systems that they have the same user and group ID number everywhere. You can edit/etc/passwd and/etc/groups with a text editor to accomplish this easily enough.
If you're talking about a Unix system and a Windows system, this requires an extra mapping facility such as PCNFS, since Windows does not provide compatible user and group ids (small integers) at all.
Neither does OS/2. What its Java-based setup tool does, however, is allow you to specify the user ID/group ID that should be associated with a directory at share time. The mount tool allows you to specify the user/group ID to use during the mount process.
Yes, it's an extra minor thing to be aware of, but I haven't met too many people who can't handle ensuring that a set of integers (x, y) are identical on all systems they need to use the share for:).
Yep and how many just download the free copy? Selling free software is a joke.
Who ever said you have to offer a copy of a GPLed program for download, online, for free?
The _only_ requirement is that you make the source available to people you distribute the binaries to. You don't have to make _any_ of it available online if you don't want to.
Sony's PS2 Linux distro is a good case-in-point. The only way to get a working copy of PS2 Linux is to buy Sony's hardware kit for around $200 USD. The DVDs that come in the kit are encoded with PS2 copy protection, so you can't readily make usuable duplicates. The source comes on DVD #2 -- and that's the _only_ way to get the kits sources from Sony. They don't provide them online for download (although, of course, as they didn't create everything "Linux" you can go around to all the websites where _they_ got the sources to collect the code as it was before Sony made any of their own modifications).
Now for code that is GPLed on those CDs, I can redistribute them myself if I wanted (although you'd be hard pressed to make them work on a PS2 without the additional hardware or the ability to press discs with the copy protection schemes intact).
Indeed, it's perfectly allowable under the GPL for me to sell you a binary-only program, licensed under the GPL, and require you to request the sources in writing, via regular letter mail, and for me to send you those sources upon request on diskettes (technically I could also send you a print-out of the source code, and require you to type it all in again, but that sort-of defeats the spirit of the license).
I'm curious as to why the story author thinks that NFS is so difficult and complicated. As I see it, from the set of systems he's listed, the only OS that should give him any trouble here would be Windows, simply because it doesn't have NFS built-in.
Here I'm running four systems -- three Intel boxes running Linux, Linux, and OS/2, and a mipsel (PS2) running Linux.
When I started setting these systems up on my network, I went through a similar investigation to interconnect all the systems for file sharing. All supported NetBIOS based sharing, and all supported TCP/IP (and had the necessary NFS client/server software).
In the end I chose NFS as it didn't require me to be running a second protocol stack on all the systems, eating up more RAM (RAM may be cheap, but I still attempt not to waste it with stuff that doesn't need to be loaded).
Here, NFS has been easy, even on the OS/2 system. NFS mounts on it get a drive letter. Specifying what to share is done via a Java application. Setup on the Linux systems is also easy -- modify a file in/etc, run a command, and then mount/unmount as necessary.
I don't know about MacOS or Windows -- I imagine that Windows might be a problem, but otherwise I think it's incorrect to throw NFS away out-of-hand as "impossibly complex", because in OSs where it's implemented correctly, it's exceedingly easy to setup and use.
There we have to agree to disagree. That is a My way or the highway attitude, which to me is force all be it a slight force.
It's no different with any other license -- that's the _whole_point_ of a license. Microsoft's license isn't any different. Nor are IBM's or Oracles or whomever elses. If you want to be completely free of restrictions, you can release your code to the public domain _without_ copyright.
Again, I'm not forcing anyone to _do_ anything. If they don't like my software license, they can use an equivilent with the license of their choosing. And if one doesn't exist, they can write their own. People who _choose_ to use my software _choose_ to abide by the terms of the license.
"Force" is diametrically opposed to te concept of "choice".
Ah yes, it's the "if I can't make a buck you can't either. situation.":-)
Absolutely not. The GPL doesn't prevent people from charging money for such a project. If you want to, you can go, right now, download the jSyncManager, and sell it to whomever you want, for however much money you want. The only requirement on you as a redistributor is that you provide the source to anyone who asks. RedHat and every other commercial Linux distribution does just this.
Well If that is the case then can Microsoft contact you, the copyright holder, and make a private deal based on new code? I.E. you grant them a different license because as the orignal copyright holder you can issue other licenses?
Certainly. All that you have to ensure is that you either own the full copyright to every bit of code you're licensing to them, or ensure that any other copyright holders have also given their permission, _or_ ensure that any code you don't own the copyright to and don't have a license to re-license (under a different license) isn't included.
Ok well that is a slight plus in favor of the GPL. I personally don't think the GPL promotes freedom. One has to have the right to use a program anyway he wants. REQUIRING anyone to do anything is not freedom it is force. But RMS is a communist and that is the way they think. Personally if I were a programer and I was inclined to do open source then I'd use BSD style. It's honest because you have to site your sources and it is free because you can do anything you want with the code.
There isn't anything to really force anyone to do anything. If someone doesn't like the way I license my software, they are simply free to not use it.
Much in the same way I don't like the way Microsoft's EULA forces me to use their code in the way they see fit. I don't like it, so I don't use their software.
In this sense, the GPL isn't really any different than any other license in a capitalist world. If you don't like it, don't use it.
From my perspective, the GPL isn't about protecting the freedom of people who want to steal my copyrighted code. It does' however, protect _my_ freedom, and the freedom of my end-users. I don't mind if someone wants to use my code in their own projects -- but in order to do so, they have to likewise release their changes to their customers (for the LGPL'd parts), or release their code to their customers (if they integrate the GPL'd parts). Again, if a developer doesn't like it, they don't have to use my code.
What I _do_ want to prevent is someone (like Microsoft) from taking my hard work, and using it to make _their_ products work, without having to give anything back to me. This is the difficulty you run into with the BSD-style licenses. For some things, this can work quite well (Ogg Vorbis is BSD-style licensed, for example, as they want their standard adopted by the widest possible audience. In this sense, what they're really marketing is their format and algorithm, with their encoding/decoding software simply being the implementational detail), but for others, you've not only given everyone the rights to your recipe, you've basically given them the ability to use that recipe to their own advantage, without having to give anything back.
There are freedoms involved with the GPL, but much like westernized democracies, just because you have rights and freedoms doesn't mean you have the right or freedom to do whatever you want (you presumably live in a "free" society, but you don't have the freedom to kill your neighbour if their grass clippings blow onto your driveway).
In the same way, I grant people a whole lot of freedoms when it comes to the jSyncManager -- more than most commercial applications do. But that doesn't mean you can do whatever you want with it (as I do have my own interests to protect too).
Ok, so why are you not using Windows or MacOSX? Because in essence that is what you are arguing. People are using LINUX and other pieces of the software because they are TOO cheap to pay for Windows.
But then how do companies such as RedHat, which charge for their Linux distros, make any money?
People _do_ pay for Linux distros on CD-ROM to alleviate them from having to download the ISOs and burn a CD set themselves.
Based on your logic, all for-pay Linux distros need not exist, because we could all go out and download the kernel, and then download all the individual programs that make a Linux distro, and setup our systems for free that way.
Well sure we _could_, but have you ever built a Linux system completely from scratch? Would you _want_ to?
Now I'll admit I'm one of those people who went the el-cheapo way and downloaded the RedHat 8 ISOs before installing them onto my system. I don't see the point in making a trip to the store to buy a boxed RedHat. However, apparently lots of people _do_ like to buy software this way, as somehow those boxed RedHat (and other Linux distros) keep selling...
I'm not a programer nor a lawyer so forgive me if I make a stupid/incorrect statemnet. Does this not depend on what license you use to begin with? If you GPL are you not required to provide the source to anyone you distributed it to? And aren't you required to GPL any changes that you distribute publicly(even if you only distribute in exchange for cash.)
Under the GPL (and pretty much every other official Open Source license) you are required to distribute the source for your program to anyone you've distributed binaries for who requests it.
That's the nature of the license -- hopefully you go into it with your eyes open, and are aware of this. If this isn't to your liking, you don't choose the GPL as your license.
Regardless, you still own the copyright on the code. Others may be able to freely use and redistribute nd even change your code, but ultimately you still own the copyrights on your own work. This is important because most countries have laws that protect copyrighted works, and none that protect license holders.
If I write program FOO 0.1 and copyright and GPL it makes changes all the way up to 0.9 and on 1.0 decide to take it to a Microsoft like EULA how does that override the requirements to GPL it. And how does that stop those who get my code from giving it away free?
As the copyright holder, you can change the license however and whenever you want. I you decide you no longer want to distribute your code under the GPL, no problem -- switch FOO 1.0 to your EULA, and the GPL no longer applies.
Note, of course, that _previous_ versions of your GPLed code may still be floating around out there, and that you'll have no real method to prevent others from using the license you've granted them to modify and redistribute your code. You can however, in effect, fork the license of your code at your whim (wheras nobody else can, as they are not the copyright holder).
As with any type of license, there are trade-offs to be made when choosing Open Source or a GPL-style license. You may not have the same sort of iron-fisted control of your code that you can have under a commercial license, but if that's what you're looking for, then Open Source isn't for you. If you don't mind giving other people the opportunity to work with your code as they see fit (within the confines of your license), then you still have ownership and control over the code -- but you've just granted others a certain amount of control over it as well, so they can better tailor it for their own use.
Lets put it this way. If I had the ability to pay nothing or a something and the difference was an installer then I would pay nothing.
Don't worry -- you don't sound like you're trolling at all. I apperciate your comments.
One thing I have learned, however, when dealing with end-users is that many of them who are _not_ skilled computer experts _will_ pay for something that is easy to setup and install.
Particularily with Java software. You'd be amazed at how many users I hear from that have a hard time with "java -jar jsyncmanager.jar". Running Java programs confuses alot of end users unnecessarily.
I wouldn't expect to sell an easily installable version to everyone who wanted to use the project. We have a nice and easy Ant-based build system, so one could argue that a sufficiently knowledgeable person doesn't even need to ever pay for binaries because they can download and build the entire project themselves.
But then again, I didn't pay anything for RedHat 8, and yet it's running on two of our project systems here:). And yet there are people who will go out and buy it in a store for the _convienence_ of it.
So I guess the idea with yours would be how to sell services or added value. For example, your sync software could provide a service where all packages to be installed are available. EG you could host X thousand projects that are referenced by your sync software. For a small monthly fee or one off fee I could use your service to install applications.
Another possible idea for our project is to sell custom jConduit plug-in development services. While we do bundle some plug-ins of generic usefulness (Installer for installing PRC/PDB files, Email for POP/SMTP based e-mail synchronization, backup and restore facilities, etc.), most organizations are probably going to want to develop jConduit plug-ins that are specific to their backend database or custom application. There may be money to be made in writing such custom add-ons (and possibly for physical transport plug-ins, which we also have an API for).
The power of free software is that so much of it's available. The free software developer can now meet just about any client needs. There's plenty of money to be made with it, and while you are making money off other people's labor why would you ever want to hoard your own? Only bad laws can stop free software from filling every software need there is. Enjoy it, take what you need and give what you can. Ideas work that way.
It's been interesting how many people have commented straight to the fund-raising portion of the story, and have neglected the marketing/advertising portion. For me, the "senses" are a bit inversed -- I'm personally mostly interested in finding better ways to get my project "out there" where people will see it -- making money is secondary.
Part of the "power of free software" as you define it is that if you have a completely _new_ idea, it may be hard to get noticed amongst all of the other excellent Open Source projects out there, even if you're solving something none of the others do. People tend to be "stuck in their ways" when it comes to the software paradigms they're used to, and if you've created a solution outside those paradigms, it can take quite a long time before you can build up a critical mass of users and watchers.
I'm not trying to pretend here that the jSyncManager has broken any seriously new ground mind you -- we've simply developed a data synchronization tool that runs as-is across platforms. This has value in and of itself, but at least it still falls into usage paradigms that PalmOS-based handheld users are used to.
Even without being radically different, however, it can be hard to get noticed out there in that sea of Open Source software. We've probably had more success than many projects out there (our SF.net activity rating is generally above 80%, and I have had the honour of speaking at conferences, we have been reviewed in the press, and we've even been mentioned in an O'Reilly book), but there are also lots of projects out there that are better known than we are.
I agree with you that you should "take what it gives" -- you've summed it up quite well. But I don't mind expending time and energy to make sure more people who are potential "givers" know about our project and its goals:).
I'm interested to see what this story does to our SourceForge statistics. I'll probably post them up in my Journal here in a day or two.
To save you a bit of trouble...4th Thrusday of every month at Futures Bakery on Bloor at 7pm. Next one is July 24.
Have any idea which subway station that's closest to? I'm out in the east end, just above the Danforth, and while I do drive, if it's readily subway-accessible, I'd rather travel "the better way" than having to pay for parking...
There is no 'marketing' of the the type of OS project your describe. A Palm-Java synchronization either works or it doesn't. If it works people pick up your code and make use of it. If it doesn't work, well....
You probably won't be suprised to find that I don't completely agree with tha assessment:).
The jSyncManager provides a synchronization framework and a set of applications that can use it, but in order to actually be useful for an organization they'll need to develop (one or more) jConduit plug-in(s) to tie into their back-end daatabases and applications. This is something that I quite obviously have quite a bit of specialization in (considering I wrote the spec:) ).
My idea of "marketing", however, was more along the lines of "how do you get people interested in your project"? So far we've had some success (I've had some interesting phone calls:) ), but it could always be better.
The really complex part of the jSyncManager is in its protocol stack. There is _no_ documentation for how the Palm's built-in synchronization protocol works, and as such I've had to reverse engineer the whole thing (and it differs between USB and serial connectivity at that). This is _not_ an easy thing to get right, as many other projects have discovered.
What's in in for a company to write you a check? By releasing your code OS, they don't jave to.
Well, just to make it clear -- I'm not looking to get rich off my Open Source project. But I also wouldn't mind getting some money coming in to help offset costs. Hence my reason for submitting my article in the first place -- it's quite likely that others here have gone through this before, and some of them may have come up with some unique and interesting ways to fund their development that I hadn't considered before.
And so far, there have been alot of interesting responses:).
I have no financial reimbursement unless developing for a company (such as IBM) and I won't own what I write, since it will be in the hands of the public domain?
First off, having myself worked for IBM as a software developer, if you work for them anything you code on your own time automatically belongs to IBM. You get no consideration for doing so, but they'll happily put their name on it.
Secondly, even though I'm giving it away for free, I still (and will always) own my code. My name is in all of the copyright statements, and the code is protected by copyright legislation both in my own country (Canada) and abroad. I have all the rights anyone else does that produces copyrightable materials. I can change the license at any time.
As John correctly points out, Open Source is more like doing volunteer work. And before you point out that you don't get paid for volunteer work, most of the bigger volunteer organizations _will_ reimburst costs associated with volunteering (I was an officer of the St. John Ambulance Brigade for 6 years, and they would routinely rebate a certain amout of your travel expenses incurred in the course of your volunteer duties).
As to your references to communism -- you couldn't be more wrong. Open Source is not a single-party system. Indeed, it is democracy that is built upon volunteerism of the sort you see in the Open Source world. The addition of bowing to commercial interests in (some) domecracies is a relatively new development. Do you get paid to vote?;)
Volunteer work forms the basis of most communities in the democratic world, and Open Source isn't any different.
I've been out of university for several years now, however as I live in Toronto (Ontario, Canada) I'm in a city with three major Universities. The university I attended is only an hour away.
I can try to see if I can connect to Business students from my old stomping grounds -- might prove to be interesting:).
In the meantime, get a day job and work on your project in your off hours. That's all the funding I've needed so far.:-)
I'm working on this as well -- but not until we release a "final" v3.0 release set and all associated documentation:).
I wouldn't mind finding someone who was willing to to the "sales as a hobby" thing, but those people are difficult to find. Sales is all about making money, and people who are dollar-centric are hard to get involved in an Open Source project that gives everything away for free (or at least appears to...).
I did try the donation thing early on in the project, but that didn't work too well. The best I ever achieved was having a member of a users group I did a presentation at buy me dinner:).
I did do slightly better talking at the Wrox Wireless Developers conference in Amsterdam in 2000 -- too bad they wound up cancelling the other conferences in the series that year:). The speakers fee was generous -- I wouldn't mind getting back into the conference speaking circuit to help fund the project.
the only truely necessary cost I can think of is bandwidth. doesn't sourceforge etc. provide this free?
Our costs have admittedly been quite low -- we are using SourceForge for pretty much everything. However, I have had to pay for my domain registration (and hosting -- I don't run my own DNS, at least not one that is publically accessible as my broadband provider doesn't permit that), our mailmaps, and I bought a system that acts as our build system and backup repository.
Those costs are pretty minor. There are other things that would be quite useful if we had even a modest budget -- consumables like business cards would certainly be useful. It would also be useful to be able to afford to keep somewhat up-to-date on the different PalmOS-based handhelds out there, to ensure that our project works well with them (currently we have to rely on user testing, and we do have some holes in terms of makes and models of handhelds tested).
I'm not looking to get rich, and I'm not expecting to make any money -- but it would certainly be nice if we could raise some funds:).
An interesting idea -- but it's a bit hard to use for a Java-based project (like the jSyncManager), which doesn't require porting to different hardware or OS platforms:).
All you need is one individual to notice your project, check it out, and realize: Hey! This is really good. Then he will tell his friends who'll tell their friends and so the stone gets rolling.
I think this works well for projects that are of "general use" -- ie, things like Operating Systems, Web Browsers, E-Mail clients, compilers/interpreters -- the type of software that every user pretty much needs (or is software that developers need to create general-use software).
There is quite a bit of software out there (like the jSyncManager) which isn't really generally applicable to all PC users -- by our nature, we're only of use to PalmOS-based handheld users. And many of the individual users (particularily those running Windows) don't need our tool, because they get a free closed-source equivilent with the handheld when they buy it (at least if they're running Windows...).
That's not to say that we don't have a market -- just that our market is vastly more fragmented. With the popularity of the Palm handheld, the popularity of Java, and the number of organizations that need a single tool that will work on all their platforms, we can fill a very important niche. However, we also realize that the people needing this sort of solution probably aren't bumping into each other at the local pub and passing their find on.
For more "niche" sorts of projects like ours, word-of-mouth can help, but it can be a really slow way to get attention, unless your niche is huge (like the "niche" that Linux fills...;) ).
Yaz.
Re:Get companies who need the software involved
on
Funding Open Source?
·
· Score: 1
Of course, this isn't useful for Palm-syncing software, since I doubt there are many companies out there with slightly odd needs with respect to Palm-syncing.
Actually, you'd be quite suprised. Corporate use is one of our fastest growing user segments (at least based on the e-mail I'we been getting), for the following reasons:
Many corporations are using mixed environments (windows/Unix/Linux/OS/2/BSD/whatever), and the jSyncManager runs the same on _all_ of them, making plug-in development and synchronization application management much simpler,
Our jConduit plug-in specification makes it a whole lot easier to create platform-neutral, use/application-specefic synchronization logic plug-ins, because it's fully Object Oriented and in Java. Many corporations are already using Java for DBMS access, and doing so with the jSyncManager is easier than with any other solution for custom in-house applications,
We're Open Source, which obviously helps quite a bit,
We're not only Open Source, but we provide a full developers API. The project was designed with a bottom-up philosophy, with the specific intention to first build a world-class synchronization API, and _then_ to build application on top of it. This makes it _very_ easy for organizations to create their own data synchronization applications, bundle data synchronization facilities into their own existing applications, or even create their own application atop our protocol stack.
Thus far, corporate users that know about us have found our project exceedingly useful. It's been put into production use within many companies for these very reasons (not to mention the fact that it works as advertised:) ).
some god damn kiddie does the same thing it took you 4 years to do, in two months and gives it aways for free. Nope. Hobbists always win.
Having once been one of those kiddies, I can say that this only really applies to the most simple applications.
The jSyncManager looks deciptively simple on the outside (on purpose -- we have a very good GUI developer who takes HCI stuff pretty seriously involved in the project), but it's inner workings are more complicated. Its multi-threaded, object oriented protocol stack is beyond even the most talented "kiddie".
It's an exceedingly difficult thing to "get right", in part because no official protocol specs have ever been published.
Anyhow, for us that has never been a fear or concern. A bigger concern is that at some point in the future Palm will simply drop their synchronization protocol altogether, or make it significantly more difficult for us to reverse engineer.
Yaz.
Re:Keep your original goals and objectives in mind
on
Funding Open Source?
·
· Score: 2, Interesting
These are exactly the sorts of little costs that have currently cropped up -- internet costs.
We've done well for some of these things -- we're using the webspace SourceForge provides for our project, for example. But domain registration still costs, as does our mailmap system.
And the project has become big enough that I can quite easily spend an entire 8+ hour day working on it, and I frequently do. As the lead developer and the project administrator and the primary project evangelist and primary technical support person, I spend my days replying to e-mails, providing end-user technical support, writing press releases, writing code, debugging code, trying to advertise our project, writing documentation -- etc.
It's all a big job. Thankfully, I have a good team of developers that has helped take some of the pressure off. Our build manager does a fantastic job, as does our webmaster (my efforts at both of these jobs in the earlier days was terrible:) ). My backup project administrator has done a fantastic job of managing bug and feature requests, and one of our testers/plug-in developers who knows CVS better than anyone else has been invaluable in helping me whenever I run into something odd with CVS (and has been a huge influence in how our CVS tree is currently organized, with more to be done after our next GA release).
But that still leaves a ton of work to do. And we're at that jumping-off stage where we're building up more and more "true believer"-type users, who are using the project in some serious projects of their own, and I'd hate to be forced to _decrease_ my participation (at least before our next GA release) because of finances.
Hire a good business person with experience raising money.
I'd love to -- know any that work for free?;)
This is part of our problem -- we have no capital to really start with. My own personal nest-egg has been dwindling over the past year just paying for rent, food, and broadband internet.
I'd love to hire a whole bunch of experts like this -- but in the current industry climate, I believe it might be difficult to find such a person who is willing to take a risk with a start-up that gives away the recipe for free:).
Slashdot Mirror
Anyway, I think Yaztromo just solved his problem #1 - by getting his question posted on Slashdot, I don't think he'll have any trouble marketing his project now, assuming it's any good.
Just FYI, now that the SourceForge statistics for July 14th (the date this story was posted to /.) are available, I decided to look them up. According to them, our homepage views were a grand total of 83 for the day.
Perhaps the results for the first full day the article was online (July 15th) will be different, but thus far having this question posted to Slashdot hasn't caused the "slashdot effect" to occur to our website at http://www.jsyncmanager.org.
I'm not disappointed by any means -- getting my question posted wasn't intended to be a marketing gimmick in and of itself. Maybe the results for yesterday, the 15th, will be more suprising (one way or another :) ).
Yaz.
a 5 line shell script will take care of getting any libraries they need from the web. There is no reason your users should have to surf to get the libs. Get them for them.
That works well for Linux users, but not so well for OS/2 or Windows users.
Part of the difficulty is also that some of these libraries have a native component, and need to be installed correctly. On Linux, some of them need to be installed as root. Additionally, at least one library (the Java Communications API for OS/2) requires a web password to download it.
Having a shell script that just downloads everything is a great idea if we were developing Linux-only software, but we're not. That shell script isn't going to run correctly outside the Unix arena, and the facilities you (and I) take for granted on Linux to make this sort of thing happen don't necessarily exist on these other platforms (and if they do, are another seperate download).
I do thank you for the otherwise good idea (I'm not trying to put it down!), but I don't see it working terribly well here for all our potential users.
Yaz.
If you run out moneys then you can go live in friendlier country based on socialism or communism instead of rude arrogant capitalist system which force people to work to get what they want.
Just ask any other Americans.
Are there any Americans in here who can tell me what this guy is talking about?
I don't know what he's thinking, but I do live in a friendlier country that is socially liberal and fiscally conservative -- I'm Canadian.
I hope he didn't just automatically assume I'm American just because I'm posting to US-centric ./ :).
Yaz.
Not to put words in the poster's mouth, but the reason many people find NFS to be somewhat complex is that the protocol requires user and group ids to be synchronized between the client and the server. If you're talking about two Unix systems, this generally means running NIS, NIS+, or LDAP, each of which is more complex than a standalone /etc/passwd file.
Nonsense I say! :). You don't need to run any extra special services like this for a small home installation -- just make sure that if a user exists on all your systems that they have the same user and group ID number everywhere. You can edit /etc/passwd and /etc/groups with a text editor to accomplish this easily enough.
If you're talking about a Unix system and a Windows system, this requires an extra mapping facility such as PCNFS, since Windows does not provide compatible user and group ids (small integers) at all.
Neither does OS/2. What its Java-based setup tool does, however, is allow you to specify the user ID/group ID that should be associated with a directory at share time. The mount tool allows you to specify the user/group ID to use during the mount process.
Yes, it's an extra minor thing to be aware of, but I haven't met too many people who can't handle ensuring that a set of integers (x, y) are identical on all systems they need to use the share for :).
Yaz.
Yep and how many just download the free copy? Selling free software is a joke.
Who ever said you have to offer a copy of a GPLed program for download, online, for free?
The _only_ requirement is that you make the source available to people you distribute the binaries to. You don't have to make _any_ of it available online if you don't want to.
Sony's PS2 Linux distro is a good case-in-point. The only way to get a working copy of PS2 Linux is to buy Sony's hardware kit for around $200 USD. The DVDs that come in the kit are encoded with PS2 copy protection, so you can't readily make usuable duplicates. The source comes on DVD #2 -- and that's the _only_ way to get the kits sources from Sony. They don't provide them online for download (although, of course, as they didn't create everything "Linux" you can go around to all the websites where _they_ got the sources to collect the code as it was before Sony made any of their own modifications).
Now for code that is GPLed on those CDs, I can redistribute them myself if I wanted (although you'd be hard pressed to make them work on a PS2 without the additional hardware or the ability to press discs with the copy protection schemes intact).
Indeed, it's perfectly allowable under the GPL for me to sell you a binary-only program, licensed under the GPL, and require you to request the sources in writing, via regular letter mail, and for me to send you those sources upon request on diskettes (technically I could also send you a print-out of the source code, and require you to type it all in again, but that sort-of defeats the spirit of the license).
Yaz.
I'm curious as to why the story author thinks that NFS is so difficult and complicated. As I see it, from the set of systems he's listed, the only OS that should give him any trouble here would be Windows, simply because it doesn't have NFS built-in.
Here I'm running four systems -- three Intel boxes running Linux, Linux, and OS/2, and a mipsel (PS2) running Linux.
When I started setting these systems up on my network, I went through a similar investigation to interconnect all the systems for file sharing. All supported NetBIOS based sharing, and all supported TCP/IP (and had the necessary NFS client/server software).
In the end I chose NFS as it didn't require me to be running a second protocol stack on all the systems, eating up more RAM (RAM may be cheap, but I still attempt not to waste it with stuff that doesn't need to be loaded).
Here, NFS has been easy, even on the OS/2 system. NFS mounts on it get a drive letter. Specifying what to share is done via a Java application. Setup on the Linux systems is also easy -- modify a file in /etc, run a command, and then mount/unmount as necessary.
I don't know about MacOS or Windows -- I imagine that Windows might be a problem, but otherwise I think it's incorrect to throw NFS away out-of-hand as "impossibly complex", because in OSs where it's implemented correctly, it's exceedingly easy to setup and use.
Yaz.
There we have to agree to disagree. That is a My way or the highway attitude, which to me is force all be it a slight force.
It's no different with any other license -- that's the _whole_point_ of a license. Microsoft's license isn't any different. Nor are IBM's or Oracles or whomever elses. If you want to be completely free of restrictions, you can release your code to the public domain _without_ copyright.
Again, I'm not forcing anyone to _do_ anything. If they don't like my software license, they can use an equivilent with the license of their choosing. And if one doesn't exist, they can write their own. People who _choose_ to use my software _choose_ to abide by the terms of the license.
"Force" is diametrically opposed to te concept of "choice".
Ah yes, it's the "if I can't make a buck you can't either. situation." :-)
Absolutely not. The GPL doesn't prevent people from charging money for such a project. If you want to, you can go, right now, download the jSyncManager, and sell it to whomever you want, for however much money you want. The only requirement on you as a redistributor is that you provide the source to anyone who asks. RedHat and every other commercial Linux distribution does just this.
Well If that is the case then can Microsoft contact you, the copyright holder, and make a private deal based on new code? I.E. you grant them a different license because as the orignal copyright holder you can issue other licenses?
Certainly. All that you have to ensure is that you either own the full copyright to every bit of code you're licensing to them, or ensure that any other copyright holders have also given their permission, _or_ ensure that any code you don't own the copyright to and don't have a license to re-license (under a different license) isn't included.
Yaz.
Ok well that is a slight plus in favor of the GPL. I personally don't think the GPL promotes freedom. One has to have the right to use a program anyway he wants. REQUIRING anyone to do anything is not freedom it is force. But RMS is a communist and that is the way they think. Personally if I were a programer and I was inclined to do open source then I'd use BSD style. It's honest because you have to site your sources and it is free because you can do anything you want with the code.
There isn't anything to really force anyone to do anything. If someone doesn't like the way I license my software, they are simply free to not use it.
Much in the same way I don't like the way Microsoft's EULA forces me to use their code in the way they see fit. I don't like it, so I don't use their software.
In this sense, the GPL isn't really any different than any other license in a capitalist world. If you don't like it, don't use it.
From my perspective, the GPL isn't about protecting the freedom of people who want to steal my copyrighted code. It does' however, protect _my_ freedom, and the freedom of my end-users. I don't mind if someone wants to use my code in their own projects -- but in order to do so, they have to likewise release their changes to their customers (for the LGPL'd parts), or release their code to their customers (if they integrate the GPL'd parts). Again, if a developer doesn't like it, they don't have to use my code.
What I _do_ want to prevent is someone (like Microsoft) from taking my hard work, and using it to make _their_ products work, without having to give anything back to me. This is the difficulty you run into with the BSD-style licenses. For some things, this can work quite well (Ogg Vorbis is BSD-style licensed, for example, as they want their standard adopted by the widest possible audience. In this sense, what they're really marketing is their format and algorithm, with their encoding/decoding software simply being the implementational detail), but for others, you've not only given everyone the rights to your recipe, you've basically given them the ability to use that recipe to their own advantage, without having to give anything back.
There are freedoms involved with the GPL, but much like westernized democracies, just because you have rights and freedoms doesn't mean you have the right or freedom to do whatever you want (you presumably live in a "free" society, but you don't have the freedom to kill your neighbour if their grass clippings blow onto your driveway).
In the same way, I grant people a whole lot of freedoms when it comes to the jSyncManager -- more than most commercial applications do. But that doesn't mean you can do whatever you want with it (as I do have my own interests to protect too).
Yaz.
Ok, so why are you not using Windows or MacOSX? Because in essence that is what you are arguing. People are using LINUX and other pieces of the software because they are TOO cheap to pay for Windows.
But then how do companies such as RedHat, which charge for their Linux distros, make any money?
People _do_ pay for Linux distros on CD-ROM to alleviate them from having to download the ISOs and burn a CD set themselves.
Based on your logic, all for-pay Linux distros need not exist, because we could all go out and download the kernel, and then download all the individual programs that make a Linux distro, and setup our systems for free that way.
Well sure we _could_, but have you ever built a Linux system completely from scratch? Would you _want_ to?
Now I'll admit I'm one of those people who went the el-cheapo way and downloaded the RedHat 8 ISOs before installing them onto my system. I don't see the point in making a trip to the store to buy a boxed RedHat. However, apparently lots of people _do_ like to buy software this way, as somehow those boxed RedHat (and other Linux distros) keep selling...
Yaz.
I'm not a programer nor a lawyer so forgive me if I make a stupid/incorrect statemnet. Does this not depend on what license you use to begin with? If you GPL are you not required to provide the source to anyone you distributed it to? And aren't you required to GPL any changes that you distribute publicly(even if you only distribute in exchange for cash.)
Under the GPL (and pretty much every other official Open Source license) you are required to distribute the source for your program to anyone you've distributed binaries for who requests it.
That's the nature of the license -- hopefully you go into it with your eyes open, and are aware of this. If this isn't to your liking, you don't choose the GPL as your license.
Regardless, you still own the copyright on the code. Others may be able to freely use and redistribute nd even change your code, but ultimately you still own the copyrights on your own work. This is important because most countries have laws that protect copyrighted works, and none that protect license holders.
If I write program FOO 0.1 and copyright and GPL it makes changes all the way up to 0.9 and on 1.0 decide to take it to a Microsoft like EULA how does that override the requirements to GPL it. And how does that stop those who get my code from giving it away free?
As the copyright holder, you can change the license however and whenever you want. I you decide you no longer want to distribute your code under the GPL, no problem -- switch FOO 1.0 to your EULA, and the GPL no longer applies.
Note, of course, that _previous_ versions of your GPLed code may still be floating around out there, and that you'll have no real method to prevent others from using the license you've granted them to modify and redistribute your code. You can however, in effect, fork the license of your code at your whim (wheras nobody else can, as they are not the copyright holder).
As with any type of license, there are trade-offs to be made when choosing Open Source or a GPL-style license. You may not have the same sort of iron-fisted control of your code that you can have under a commercial license, but if that's what you're looking for, then Open Source isn't for you. If you don't mind giving other people the opportunity to work with your code as they see fit (within the confines of your license), then you still have ownership and control over the code -- but you've just granted others a certain amount of control over it as well, so they can better tailor it for their own use.
HTH!
Yaz.
Lets put it this way. If I had the ability to pay nothing or a something and the difference was an installer then I would pay nothing.
Don't worry -- you don't sound like you're trolling at all. I apperciate your comments.
One thing I have learned, however, when dealing with end-users is that many of them who are _not_ skilled computer experts _will_ pay for something that is easy to setup and install.
Particularily with Java software. You'd be amazed at how many users I hear from that have a hard time with "java -jar jsyncmanager.jar". Running Java programs confuses alot of end users unnecessarily.
I wouldn't expect to sell an easily installable version to everyone who wanted to use the project. We have a nice and easy Ant-based build system, so one could argue that a sufficiently knowledgeable person doesn't even need to ever pay for binaries because they can download and build the entire project themselves.
But then again, I didn't pay anything for RedHat 8, and yet it's running on two of our project systems here :). And yet there are people who will go out and buy it in a store for the _convienence_ of it.
So I guess the idea with yours would be how to sell services or added value. For example, your sync software could provide a service where all packages to be installed are available. EG you could host X thousand projects that are referenced by your sync software. For a small monthly fee or one off fee I could use your service to install applications.
Another possible idea for our project is to sell custom jConduit plug-in development services. While we do bundle some plug-ins of generic usefulness (Installer for installing PRC/PDB files, Email for POP/SMTP based e-mail synchronization, backup and restore facilities, etc.), most organizations are probably going to want to develop jConduit plug-ins that are specific to their backend database or custom application. There may be money to be made in writing such custom add-ons (and possibly for physical transport plug-ins, which we also have an API for).
Thanks for your input!
Yaz.
The power of free software is that so much of it's available. The free software developer can now meet just about any client needs. There's plenty of money to be made with it, and while you are making money off other people's labor why would you ever want to hoard your own? Only bad laws can stop free software from filling every software need there is. Enjoy it, take what you need and give what you can. Ideas work that way.
It's been interesting how many people have commented straight to the fund-raising portion of the story, and have neglected the marketing/advertising portion. For me, the "senses" are a bit inversed -- I'm personally mostly interested in finding better ways to get my project "out there" where people will see it -- making money is secondary.
Part of the "power of free software" as you define it is that if you have a completely _new_ idea, it may be hard to get noticed amongst all of the other excellent Open Source projects out there, even if you're solving something none of the others do. People tend to be "stuck in their ways" when it comes to the software paradigms they're used to, and if you've created a solution outside those paradigms, it can take quite a long time before you can build up a critical mass of users and watchers.
I'm not trying to pretend here that the jSyncManager has broken any seriously new ground mind you -- we've simply developed a data synchronization tool that runs as-is across platforms. This has value in and of itself, but at least it still falls into usage paradigms that PalmOS-based handheld users are used to.
Even without being radically different, however, it can be hard to get noticed out there in that sea of Open Source software. We've probably had more success than many projects out there (our SF.net activity rating is generally above 80%, and I have had the honour of speaking at conferences, we have been reviewed in the press, and we've even been mentioned in an O'Reilly book), but there are also lots of projects out there that are better known than we are.
I agree with you that you should "take what it gives" -- you've summed it up quite well. But I don't mind expending time and energy to make sure more people who are potential "givers" know about our project and its goals :).
I'm interested to see what this story does to our SourceForge statistics. I'll probably post them up in my Journal here in a day or two.
Yaz.
To save you a bit of trouble...4th Thrusday of every month at Futures Bakery on Bloor at 7pm. Next one is July 24.
Have any idea which subway station that's closest to? I'm out in the east end, just above the Danforth, and while I do drive, if it's readily subway-accessible, I'd rather travel "the better way" than having to pay for parking...
Thanks!
Yaz.
There is no 'marketing' of the the type of OS project your describe. A Palm-Java synchronization either works or it doesn't. If it works people pick up your code and make use of it. If it doesn't work, well....
You probably won't be suprised to find that I don't completely agree with tha assessment :).
The jSyncManager provides a synchronization framework and a set of applications that can use it, but in order to actually be useful for an organization they'll need to develop (one or more) jConduit plug-in(s) to tie into their back-end daatabases and applications. This is something that I quite obviously have quite a bit of specialization in (considering I wrote the spec :) ).
My idea of "marketing", however, was more along the lines of "how do you get people interested in your project"? So far we've had some success (I've had some interesting phone calls :) ), but it could always be better.
The really complex part of the jSyncManager is in its protocol stack. There is _no_ documentation for how the Palm's built-in synchronization protocol works, and as such I've had to reverse engineer the whole thing (and it differs between USB and serial connectivity at that). This is _not_ an easy thing to get right, as many other projects have discovered.
What's in in for a company to write you a check? By releasing your code OS, they don't jave to.
Well, just to make it clear -- I'm not looking to get rich off my Open Source project. But I also wouldn't mind getting some money coming in to help offset costs. Hence my reason for submitting my article in the first place -- it's quite likely that others here have gone through this before, and some of them may have come up with some unique and interesting ways to fund their development that I hadn't considered before.
And so far, there have been alot of interesting responses :).
Yaz.
I have no financial reimbursement unless developing for a company (such as IBM) and I won't own what I write, since it will be in the hands of the public domain?
First off, having myself worked for IBM as a software developer, if you work for them anything you code on your own time automatically belongs to IBM. You get no consideration for doing so, but they'll happily put their name on it.
Secondly, even though I'm giving it away for free, I still (and will always) own my code. My name is in all of the copyright statements, and the code is protected by copyright legislation both in my own country (Canada) and abroad. I have all the rights anyone else does that produces copyrightable materials. I can change the license at any time.
As John correctly points out, Open Source is more like doing volunteer work. And before you point out that you don't get paid for volunteer work, most of the bigger volunteer organizations _will_ reimburst costs associated with volunteering (I was an officer of the St. John Ambulance Brigade for 6 years, and they would routinely rebate a certain amout of your travel expenses incurred in the course of your volunteer duties).
As to your references to communism -- you couldn't be more wrong. Open Source is not a single-party system. Indeed, it is democracy that is built upon volunteerism of the sort you see in the Open Source world. The addition of bowing to commercial interests in (some) domecracies is a relatively new development. Do you get paid to vote? ;)
Volunteer work forms the basis of most communities in the democratic world, and Open Source isn't any different.
Yaz.
Are you at or near a university?
I've been out of university for several years now, however as I live in Toronto (Ontario, Canada) I'm in a city with three major Universities. The university I attended is only an hour away.
I can try to see if I can connect to Business students from my old stomping grounds -- might prove to be interesting :).
Yaz.
In the meantime, get a day job and work on your project in your off hours. That's all the funding I've needed so far. :-)
I'm working on this as well -- but not until we release a "final" v3.0 release set and all associated documentation :).
I wouldn't mind finding someone who was willing to to the "sales as a hobby" thing, but those people are difficult to find. Sales is all about making money, and people who are dollar-centric are hard to get involved in an Open Source project that gives everything away for free (or at least appears to...).
I did try the donation thing early on in the project, but that didn't work too well. The best I ever achieved was having a member of a users group I did a presentation at buy me dinner :).
I did do slightly better talking at the Wrox Wireless Developers conference in Amsterdam in 2000 -- too bad they wound up cancelling the other conferences in the series that year :). The speakers fee was generous -- I wouldn't mind getting back into the conference speaking circuit to help fund the project.
Yaz.
The jSyncManager Project is actually based in Toronto -- when/where do these meetings occur? Are they posted somewhere?
Yaz.
the only truely necessary cost I can think of is bandwidth. doesn't sourceforge etc. provide this free?
Our costs have admittedly been quite low -- we are using SourceForge for pretty much everything. However, I have had to pay for my domain registration (and hosting -- I don't run my own DNS, at least not one that is publically accessible as my broadband provider doesn't permit that), our mailmaps, and I bought a system that acts as our build system and backup repository.
Those costs are pretty minor. There are other things that would be quite useful if we had even a modest budget -- consumables like business cards would certainly be useful. It would also be useful to be able to afford to keep somewhat up-to-date on the different PalmOS-based handhelds out there, to ensure that our project works well with them (currently we have to rely on user testing, and we do have some holes in terms of makes and models of handhelds tested).
I'm not looking to get rich, and I'm not expecting to make any money -- but it would certainly be nice if we could raise some funds :).
Yaz.
An interesting idea -- but it's a bit hard to use for a Java-based project (like the jSyncManager), which doesn't require porting to different hardware or OS platforms :).
Yaz.
All you need is one individual to notice your project, check it out, and realize: Hey! This is really good. Then he will tell his friends who'll tell their friends and so the stone gets rolling.
I think this works well for projects that are of "general use" -- ie, things like Operating Systems, Web Browsers, E-Mail clients, compilers/interpreters -- the type of software that every user pretty much needs (or is software that developers need to create general-use software).
There is quite a bit of software out there (like the jSyncManager) which isn't really generally applicable to all PC users -- by our nature, we're only of use to PalmOS-based handheld users. And many of the individual users (particularily those running Windows) don't need our tool, because they get a free closed-source equivilent with the handheld when they buy it (at least if they're running Windows...).
That's not to say that we don't have a market -- just that our market is vastly more fragmented. With the popularity of the Palm handheld, the popularity of Java, and the number of organizations that need a single tool that will work on all their platforms, we can fill a very important niche. However, we also realize that the people needing this sort of solution probably aren't bumping into each other at the local pub and passing their find on.
For more "niche" sorts of projects like ours, word-of-mouth can help, but it can be a really slow way to get attention, unless your niche is huge (like the "niche" that Linux fills... ;) ).
Yaz.
Of course, this isn't useful for Palm-syncing software, since I doubt there are many companies out there with slightly odd needs with respect to Palm-syncing.
Actually, you'd be quite suprised. Corporate use is one of our fastest growing user segments (at least based on the e-mail I'we been getting), for the following reasons:
Thus far, corporate users that know about us have found our project exceedingly useful. It's been put into production use within many companies for these very reasons (not to mention the fact that it works as advertised :) ).
Yaz.
some god damn kiddie does the same thing it took you 4 years to do, in two months and gives it aways for free. Nope. Hobbists always win.
Having once been one of those kiddies, I can say that this only really applies to the most simple applications.
The jSyncManager looks deciptively simple on the outside (on purpose -- we have a very good GUI developer who takes HCI stuff pretty seriously involved in the project), but it's inner workings are more complicated. Its multi-threaded, object oriented protocol stack is beyond even the most talented "kiddie".
It's an exceedingly difficult thing to "get right", in part because no official protocol specs have ever been published.
Anyhow, for us that has never been a fear or concern. A bigger concern is that at some point in the future Palm will simply drop their synchronization protocol altogether, or make it significantly more difficult for us to reverse engineer.
Yaz.
These are exactly the sorts of little costs that have currently cropped up -- internet costs.
We've done well for some of these things -- we're using the webspace SourceForge provides for our project, for example. But domain registration still costs, as does our mailmap system.
And the project has become big enough that I can quite easily spend an entire 8+ hour day working on it, and I frequently do. As the lead developer and the project administrator and the primary project evangelist and primary technical support person, I spend my days replying to e-mails, providing end-user technical support, writing press releases, writing code, debugging code, trying to advertise our project, writing documentation -- etc.
It's all a big job. Thankfully, I have a good team of developers that has helped take some of the pressure off. Our build manager does a fantastic job, as does our webmaster (my efforts at both of these jobs in the earlier days was terrible :) ). My backup project administrator has done a fantastic job of managing bug and feature requests, and one of our testers/plug-in developers who knows CVS better than anyone else has been invaluable in helping me whenever I run into something odd with CVS (and has been a huge influence in how our CVS tree is currently organized, with more to be done after our next GA release).
But that still leaves a ton of work to do. And we're at that jumping-off stage where we're building up more and more "true believer"-type users, who are using the project in some serious projects of their own, and I'd hate to be forced to _decrease_ my participation (at least before our next GA release) because of finances.
Yaz.
Hire a good business person with experience raising money.
I'd love to -- know any that work for free? ;)
This is part of our problem -- we have no capital to really start with. My own personal nest-egg has been dwindling over the past year just paying for rent, food, and broadband internet.
I'd love to hire a whole bunch of experts like this -- but in the current industry climate, I believe it might be difficult to find such a person who is willing to take a risk with a start-up that gives away the recipe for free :).
Yaz.