Anatomy of an Open-Source Project

[R.Daugherty]

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
I would think this group which generally advocates a web browser interface would ask themselves why IBM created a new open source Eclipse desktop client for Notes if I understand the marketing announcement correctly. Apparently I didn't understand the announcement correctly. IBM talked of open source Eclipse and an open client based on Eclipse, but not an open source client as I posted. What makes IBM's desktop client open when it's not open source, and what makes other desktop clients without source code closed in comparison to IBM's? IBM says one avoids lockin with their open client. How, if it's not open source? What relevance does Eclipse being open source have to an IBM product based on it that isn't? I don't know, probably about as much relevance as me thinking it could be applied as the iseries interface. rd

[rward@mcind.com]

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
I recently ran across a product called Morfik (www.morfik.com). The project is young but shows a lot of promise. Adobe's Flex and OpenLaslo are also projects we've been looking at as a possible way to bring rich GUIs to our Internet users. As stated before, the current state of affairs for web application programming is way too complex. To build functional, professional web applications today you have to know HTML, DHTML, XHTML, CSS, Javascript, XML, PHP,JSP, ASP, PHP, toolkit flavor of the day, template engines, frameworks, etc, etc, etc.

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
It is certainly interesting times. What we do need is a Web programming environment that is not incremental but tangential; something like when VB3 was introduced in the micro world. Such a product will be a turning point. Of course, it may never happen.

[R.Daugherty]

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
You sum up the general consensus well. I could quibble, but I won't argue with any of it. I do agree with Thomas' point of innovation with open source, however. Even with commercial software the innovations are incremental these days. There will be another plateau reached one day, with combinatorial interaction of software enabled through today's work of open source interface architectures, but critical mass for that has not been reached yet. I took a look at Oracle's JD Edwards white paper on the newly released A9.1. I note that they added function key support to the web interface, and optional Java replacements for web pages, both at users requests. That the users are iseries veteran users, and know what they want, is one factor in that. They would not have asked if they found web pages a replacement for 5250. SAP would not be using Flash if they found web pages a replacement for Unix screens. Microsoft would not be inserting Windows into the browser with ActiveX and WebForms if users found web pages a replacement for Windows screens. IBM would not have developed an Eclipse client for Notes if users found web pages a replacement for Notes. All I want is a standard interface for the iseries that is as good as what other vendors are smart enough to provide their customers. Every app having its own crazy quilt web secret sauce going, with IBM trying to lasso everyone into corralling them through a Websphere tollgate, is not viable. Until then, I guess we'll see if stateless web architectures can compete with 5250 resident sessions. We could do more. Oracle did with JDE World. That I find the height of irony. rd

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
Fundamentally though, both 5250 and Web interfaces are form processors that was my point. When you compare the 3 environments under discussion, 5250 and Web interfaces are so much more identical than the thick client environments. There is no question Windows and Linux (eg. Qt) offer a superior event driven programming environment but at a cost. Now 5250 is a document handler that has indeed been shoe-horned to a form processing role. But it is ubiquitous. It is the lingual franca of the world at the moment. Most devices have a host environment. Computers (mainframe, midrange, servers, workstations, laptops), mobile devices, game consoles and even appliance equipment. 5250 streams can be hosted only on certain computing platforms and then it requires it's own special host environment and controls to operate. On the thick client implementation it also requires numerous TCP ports to operate. Whilst this can be seen as an asset is it actually a hinderance to a lot of operation managers who are trying to tie down security. A problem with HTML use is the implementation. I think David refers to this in a following post. It's really a cross between the expediency of using a commonly accessible platform (web servers, browsers), the desire of the programming rituous with a bent of layer separation at all costs, and the need for scalability on the smallest possible piece of hardware to a do a task (a big reason for Stateless programming). Ajax and the like are hacks but they are filling gaps in usability. As usual, overuse kills the bird but there are many good implementations. From and ISV point of view, 5250 is a dead event and thick windows client apps are fast heading that way. You have a proliferation of middle tier app servers popping up everywhere (Citrix, Sun One, et al). They are clearly the stepping stones to what a lot of IS are chasing for the future. That is why I wonder at the wisdom (from the original thread topic) of pushing Open Source as an alternative when, with any qualitative analysis, it is merely acting as a free replacement of existing functionality. And worse still, functionality that may not be required too far in the future. At no point is there any real innovation - it is more about copying existing ideas or covering existing ground.

[David Abramowitz]

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
Ralph Daugherty wrote: Obviously we could use better font, color, and layout control, but even there the same thing can be said about a browser, which was never intended to have the layout control asked of it. It's kludge upon kludge to have what has been done with it, remarkable as it is. IMHO - Work on the 5250 interface was left incomplete. There is much more that could be done, even within the existing confines. IBM has pretty much chosen to leave it alone, and that, is a pity. Browser interface programming has diversified to the point of incomprehensibility. Ads for web developers insist on all of the following: HTML, DHTML, XHTML, ASP, JSP, VB script, Javascript, PHP, AJAX, XML, CSS, and the litany goes on and on and on. A single shop typically has purchased multiple programs each with its own technology to achieve a similar goal through the browser. Keeping up is like chasing a moving target. As soon as you think you have mastered one method, either the method changes, or is outdated by the latest programming flavor of the month. Dave

[R.Daugherty]

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
The selection lists with 5250, cursor sensitive subfile windows, are much more powerful than drop down lists which in lieu of AJAX type programming must all be sent down with the web page to display in the drop down window. But similarities of the interface are the least of it. It is the "stream" processing which in fact is not comparable at all. 5250 is sockets communications to a resident session. HTML is, well, it's a kludge pushed into doing things it was never designed to do. That's why the security nightmare known as ActiveX was used and required Windows and Internet Explorer for a so called access from anywhere solution known as the browser. Now Flash is used instead most of the time. Flash is a native program running in a browser. SAP is going to it for an interface, for example. So anyone doing less is just behind the eight ball to start with. AJAX just compounds that kludge infinitely. Does anyone have any idea what must take place on the server every time an AJAX request is sent? A similar situation would exist if 5250 went to keystroke processing. With careful design AJAX requests could occur no more frequently than a page refresh and bring down less data, and the same could be said of 5250 keystroke processing. But these type of processing requests tend to generate lots of request traffic to the server for most purposes. So no, unfortunately for web browser advocates, there is no comparison to the stream processing of 5250 and HTML, and 5250 has more powerful interface processing than a web browser. Obviously we could use better font, color, and layout control, but even there the same thing can be said about a browser, which was never intended to have the layout control asked of it. It's kludge upon kludge to have what has been done with it, remarkable as it is. rd

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
From our experience the problem with Windows on the client is not licensing costs it is the support demand for the environment; particularly when new programs are added or upgraded with their associated COM baggage and shortcuts. Due to security restrictions these steps have a need to have a user with local admin rights to perform the tasks. In a lot of sites it is still common to see an individual run around to each workstation to install or upgrade them. If the IS services can deploy apps from Web or Citrix servers then all this management is pushed to a back end server. The Windows environment becomes significantly cleaner and easier to manage as they move more to a dumb terminal environment. This doesn't mean that ALL sessions are easily moved this way. There are advantages for some apps, eg program development environments, to be locally hosted on a workstation. However there is a counter point. Some Db vendors went keen on Java for instance for their DB tools as a "Universal" client solution. Good idea but you can see the distinct speed difference between these apps over native apps from other vendors on Windows clients. Both work though if you think broader. There is an argument that 5250 sessions are somehow more advanced than Web sessions. But in reality they are very similar - a stream being sent to the client to be interpreted and presented by the Terminal or host app (IBM Access or IE). Both environments have just form based processing at heart. Fill the form - push ENTER. HTML does support a lot greater extension of form controls (drop down boxes, selection lists, radio buttons, etc). I think in the past the clunkiness of the Web interface was it's undoing. We also had similar issues with the 5250 form processing orientation vs the more fluid thick client interfaces where messages are presented dynamically or new form fields appear dynamically when a checkbox is clicked etc. But even here the Web interfaces are moving ahead with Ajax processing that enables HTML to look more and more like thick client sessions.

[R.Daugherty]

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
As a extension to the desktop argument above - if businesses are trying to get away from Windows Clients why would they want the Linux Clones? I don't think it's valid at all that business is trying to get away from Windows. I think they would move reluctantly if costs and licensing and security issues are what they are and Microsoft tries to force them to Vista. Run desktop software via browser? There is a recent announcement of something along those lines of a so called web offering from a venture for Open Office, but even that uses a Windows client, not a browser. There is a recent slashdot thread on that, and many comments on desktop apps via browser, none of them positive. I would think this group which generally advocates a web browser interface would ask themselves why IBM created a new open source Eclipse desktop client for Notes if I understand the marketing announcement correctly. That is the kind of interface that will run on any desktop OS or thin client. That's just one example of the kind of open source solutions that I'm thinking of. And one last thought. Why is it that high powered software like games run with sockets instead of browsers? Isn't that the kind of power winning business software will bring to the table? That and a mixture of other collaborative desktop tools like spreadsheets and other productivity software? So if Notes and games need high powered desktop interfaces, why is not a browser wimping out on the solutions we could provide our customers? And an open source infrastructure solution, including IBM's desktop client for example. rd

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
"There used to be an old saying that nobody ever got fired for recommending IBM. Replace IBM with Microsoft for today's environment." That is why I think MS is also living on borrowed time. Their key "marketting" players are aging and I see their dominance possibly diminishing within the next 10 years. Certainly the desktop environment is the current major seller yet also represents the major flaw in their empire. Even now businesses are looking for less complexity not more on the client environment. Hence the push for Web Serving and Citrix/Terminal Server. As a extension to the desktop argument above - if businesses are trying to get away from Windows Clients why would they want the Linux Clones? Any new dominant player will NOT be Open Source. This model does have a commercial imperative and will never dominate. You are looking at a pseudo "socialist" vs "capitalist" fight again. Open Source may have a ride on the current generation of contributors but the next may not wish to invest their time into the Open Source, will see others in the non Open Source industry doing better than themselves, or will perhaps even take a completely different career due to diminished financial return on general programming effort. If there is a new player I see that it will come from the non open source software development industry. It will certainly not be from any hardware industry player where they are happy to service whatever is on offer (eg. MS, Unix, Linux). It will have a heavy focus on Web operating environments and the mobile industry. This would be more like the MS course to dominance rather than the IBM course (which in the 50s was basically hardware oriented).

[R.Daugherty]

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
They're out there selling something that isn't open source. I think the Microsoft solution is expensive, both up front and licensing wise, but the businesses must feel like they're getting their money's worth. A similar situation exists with IBM and those of us on iseries feel there is justifiable cost effectiveness. But will anyone else ever feel that way about it, and will people resist Vista? Whatever the answer is, it may be Sun Solaris and Java made easier, or Linux and C++ made easier, it will be an open source answer. rd

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
Ralph, I agree with your view as reality. Open Office couldn't compete as a cheaper commercial alternative to MS Office. It still can't as a freebie. I use Open Office and love it but trying to get businesses to replace MS Office is a very tough sell. I've been marketing myself locally as an open source consultant and have been getting very few hits. There used to be an old saying that nobody ever got fired for recommending IBM. Replace IBM with Microsoft for today's environment. Very few people want to replace Office with Open Office. Very few people want their websites developed in PHP and MySQL. They want ASP.NET and SQL Server. I have to pay the mortgage, so I stick with what pays. Some things I've learned along the way working with open source products: 1. Open source does not mean free. 2. Open source does not mean it can work with everything. 3. Open source does not mean you'll be able to or want to actually modify the source, recompile, and make your own changes. 4. Open source does not mean a lack of functionality. 5. Open source does not mean amateur, although a large number of amateurs contribute. 6. Open source does not mean Microsoft is going away any time soon. This link provides what I think is a concise description open source. It describes four types of "open source". http://www.newsforge.com/article.pl?...57205&from=rss It describes the motives for licensing under an open source banner. It also rails against using open source as a verb. ;}. I need to find some deep pockets... you know any Ralph? Tom.

[R.Daugherty]

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
I see your point and I need to quickly clarify. When I say that software can't sell, I did not mean to imply it was a second class citizen! I meant it can't sell, and I mentioned monopoly in an earlier post. I also advocated giving back to the open source community in my earlier post. At no point have I ever suggested that open source was not the route to go. My own plans are to see how open source can be applied to the iseries, and give back in the process. Let's take Open Office specifically. It was a commercial product that couldn't get any traction against the Office juggernaut. The Office juggernaut kept getting its formats slightly changed like a virus to keep anything from being able to lock in and participate in that ecosystem. So Open Office, no matter what quality the software, withered. Sun bought it and open sourced it. They are deep pockets with an ulterior motive. They were in serious competition with Microsoft. There was simply no choice but to open source to compete with Microsoft. Any other take more noble and pure is simply wishful thinking. Most open source work is done by deep pockets with ulterior motives. This especially includes IBM. They and other deep pockets only open source that which they need to establish a market. This is only a natural extension of market driven economics. There are some sucessful pure plays. Linux is the premier example. But all the work it takes to make it a successful competitor to Windows wasn't getting done. Enter a deep pocket from South Africa who is funding that work. I wrote in my earlier post that there are pure of heart just working to have an alternative to a monopoly, but it's easier to be pure of heart when you're not a commercial competitor. So can't sell and ulterior motives may sound antagonistic, but they are the truth, and nothing is more noble than the truth. rd

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
First of all, the perspective that open source products are somehow "second class citizens" is specious. If you take a look at some of the most popular open source products, such as Open Office or Joomla!, and really examine the level of functionality, you will discover that the level of sophistication and innovation in solving problems is quite impressive. Sure there are thousands of open source "freeware" programs that do not reach that level of sophistication. Sure there are thousands of projects that are the result of one or two developers that will probably never gain great acceptance. But to believe that an open source project like Linux is somehow inferior to Windows is ludicrous today. Secondly, innovation is not only in how a product reproduces current functions in a set, but how that set is developed and delivered. Professionals laughed at the idea that the open source movement would gain momentum. Now quality products are being developed by diverse teams -- and enhanced at an astounding rate -- much faster than commercial organizations can deliver. That, in and of itself, is an innovation. Finally, as I said in an earlier post, we are at the second phase of the open source evolution: The phase that imitates functionality of a feature set from commercial products. Just as Firefox took away Microsoft's lead in innovation for browsers, so too will highly functional features in standard business sets soon be out-evolving commercial suites. Why? Because commercial development of software is much more expensive than the open source model. Why is it more expensive? Because ten minutes of a commercial programmer's time costs hundreds of dollars more than ten minutes of ten open source programmers who are in the game as a problem-solving, career-building exercise, volunteering their energies. It's a community dynamic that will be very difficult for commercial entities to compete against. So, innovation is not only about the current release of any particular product set. It's also about how the feature set is engineered, developed, and delivered. When those elements are dynamic -- as they are in the best open source communities -- then feature innovation and ultimately product innovation will soon outstrip the ability of any single company to deliver. That's why IBM is embracing open source. They will develop middleware products that will enable open source because it's the fastest way to expand the market for software. And their strategy is to be the provider of middleware products to that market.

[R.Daugherty]

Anatomy of an Open-Source Project

** This thread discusses the article: Anatomy of an Open-Source Project **
You are right. I'm cynical, jaded, and a pain in the butt to say this, but the only innovation in open source is from those trying to establish market share that they wouldn't be able to do otherwise selling it. Almost always it's architectures. A corollary to that is the software that couldn't sell, so the next best thing to do was to open source it. Sometimes a company with deep pockets that wasn't competitive in an area buys the company and open sources it. The vast remainder are people working to develop alternatives to established monopolies. They are not innovations, they are free alternatives. Bottom line, innovative software that can sell is sold unless deep pockets have ulterior motives. The most innovative software I know of that was developed open source was funded by a deep pocket who made his fortune on software with copy protection at one point. Some of the most innovative software is funded by deep pockets and free, but little of it is open source. There is always a self interested reason when it is, except for those pure of heart trying to uninnovatively provide an alternative to a monopoly. rd