Why try ivtools?

Why should a developer of commercial or free software applications for Unix consider using ivtools for their work?
  • * mature framework for direct-manipulation graphical editors
  • * lightweight glyphs for interactive structured document applications
  • * separation of structured graphic and structured document classes
  • * integration of an extendable and distributed command interpreter
  • * freedom to use the source in any fashion
  • * protection from software patents
  • * continual commercial-quality development and support
  • * organized leverage of other software frameworks
  • * open coordination of distributed development via patch
  • * mature framework for direct-manipulation graphical editors

    ivtools is a mature framework for constructing custom drawing editors and direct manipulation user interfaces. Starting with John Vlissides' Unidraw, we've added layers of class libraries for graph and frame visualization, image display and plot importing, as well as augmenting the original software with support for alternate document formats, command interpreting, attribute (property-list) management, and networking capabilities.

    * lightweight glyphs for interactive structured document applications

    The InterViews team was among the first to explore the concept of light-weight user interface objects, where pick and layout are managed on the client side. They used a single X window for a complex glyph layout, shifting more of the rendering logic to the client. The client gets finer grained control and the X server is left to do what it does best, paint raster and vector graphics.

    This approach had been pioneered with the structured graphics capability of InterViews 2.6. Now with glyphs the same performance advantages were brought to conventional widget objects (button, slider, menus, etc..). The improvement relative to Motif for a complicated dialog box was quite noticeable, hence the term "flyweight". In addition these glyphs were put to the test in creating doc, a document editor with TeX like capabilities where C++ objects were used down to the single character level. You can read more on this here

    * separation of structured graphic and structured document classes

    ivtools (and InterViews 3.1) has separate sets of objects for structured graphics and structured douments. Fresco has a single set of classes for structured graphics and structured documents. Unification of these two different fields of applications is interesting, but in our opinion unnecessarily complicates the API and the debugging process if all you really want is a custom drawing editor. It seems either you're building a pure graphical interface, where no automated layout is required, or your building a boxes-and-glue interface where you want all the power of TeX under the hood. In ivtools these classes are still separate, and different programmers can specialize in one or the other.

    * integration of an extendable and distributed command interpreter

    The command interpreter can be used within applications built on ivtools as a calculator for computing arbitrary attribute values, for selecting objects in a drawing editor by evaluating an arbitrary attribute expression, for controlling the viewer mechanism of a drawing editor and the playback mechanism of a flipbook viewer, and even for scripted drawing of graphics as an alternate to manual interactive creation. All of these are available for reuse in custom applications.

    In their ACM paper on Unidraw, Linton and Vlissides concluded with a discussion of what could be done to further extend the usefulness of the framework. One thing they mentioned was the ability to interpret external formats, thus translating them into an internal representation that can be saved/restored, and exported to other formats. You could almost already do this with ivtools, because the language used to save/restore documents is the same one recognized by the command interpreter. A given interpreter, like that in comdraw could be expanded with a new function for each object type in the specialized application document.

    * freedom to use the source in any fashion

    ivtools is open source and non-copylefted X Consortium style free software. This means you can use it for both free-software projects and commercial software development. You can even GPL or LGPL your copy if you so desire, but we'll carry on with the license the way it is, the way it has always been *. There is no document guaranteeing this, but the logic behind Vectaport's business should be self-evident. We make money off of selling service, training, and CD-ROMs. Our customers are in the business of making custom, often proprietary, applications. They understand the advantage of leveraging a freely available foundation of generic software with few strings attached. Vectaport ceased existence as a corporation Dec 31st 2000, but continues as a web site and consulting venue for Scott Johnston

    Ok, there is a string or two. The copyright notices we used are derived from the X11 copyright, and require that you include the entire copyright and permission notice at least once, somewhere in the documentation, in any derivative work, and that at least the single copyright line itself remain in the source code. And that you not use the name of any copyright holder in any form of publicity or advertising without permission.

    * protection from software patents

    This is an interesting advantage of ivtools (and InterViews 3.1). All of the core mechanisms of the drawing editors, flyweight glyphs, and command interpreters were published in source or paper form in the late 1980s or early 1990s. With a decade head-start on the forces of patent litigation, you can feel about as safe investing in this software technology as you can hope for. Not only in your ability to win if taken to court, but in your ability to stay out of court and brush off nuisance licensors, by flooding them with a collection of long-time URL's that convincingly demonstrate prior art. The original interviews.stanford.edu ftp site is still up and running, though nothing has changed on it for years.

    * continual commercial-quality development and support

    At Vectaport we've been programming with InterViews 3.1 and derivative source code since 1993. Our experience with InterViews started with using InterViews 2.6 in 1990. And the initial development for the command interpreter embedded into ivtools was undertaken as early as 1988. Vectaport ceased existence as a corporation Dec 31st 2000, but continues as a web site and consulting venue for Scott Johnston

    Though some of the code is old, some of the code is always new. We continually develop and release these frameworks, incrementally adding capability month after month, year after year. We take care to make sure any changes to the API or implementation of existing classes are the kind that can safely be propogated by recompiling the C++ source. Unlike C API's, where people respect finished and unchanging libraries of code, these C++ frameworks are continually changing, continually growing, and continually being improved, all in a way that doesn't upset any particularily static use of the classes as they existed at some prior time. To prove this we continually seek out static InterViews applications, like MixViews, and test that they can still be built on top of ivtools.

    * organized leverage of other software frameworks

    Is a nice drawing editor not enough reason to consider using ivtools? Then what about a powerful middleware package for networking and system services, a collection of frameworks for image understanding, a sound-editing application, packages for geo-spatial map rendering in 2d and 3d and vehicle simulation for navigating those maps? All of these are software packages that work together with ivtools, software packages where we have done the initial interfacing or integration work, and published the results for others to use.

    * open coordination of distributed development via patch

    You'll find no centralized use of cvs at vectaport.com, because we appreciate the right of any programmer to manage their own copies of the source they rely on, and to incorporate the changes of third parties whenever they please. We see no need to force them to upgrade with everyone elses latest changes as a pre-requisite for committing and distributing their own changes (the cvs model of coordinating development). Instead we use cvs as a personal or local-area tool for configuration management and for generating patches for sharing with remote sites. The wizardry of patch is that it can allow disparate programmers to share improvements between source trees of similar lineage, without ever absolutely requiring the merging or unification of the source trees. If and when this is done is up to the programmers, not the tool. If a particular configuration of software is well-suited and well-tested for a well-defined application, what is wrong with it never changing again, or only changing to incorporate necessary bug fixes while ignoring any unnecessary enhancements or extensions?

    However, the converse question can also be asked. What is wrong with incorporating unnecessary enhancements and extensions if the stable application still compiles, links, and works? We've found a typical approach of third party developers is to remain with a stable release of ivtools while building their own applications, incorporating only the patches they need for the particular bugs they find or new features they need. Then, if and when they have the time, they upgrade to the latest version of ivtools.