"Linux Gazette...making Linux just a little more fun!"

 

Set your browser as wide as you'd like now.  I've fixed the Muse to expand to fill the aviailable space! © 1997 by mjh

 

muse:  v; to become absorbed in thought  n; [ fr. Any of the nine sister goddesses of learning and the arts in Greek Mythology ]: a source of inspiration   elcome to the Graphics Muse! Why a "muse"? Well, except for the sisters aspect, the above definitions are pretty much the way I'd describe my own interest in computer graphics: it keeps me deep in thought and it is a daily source of inspiration.  [Graphics Mews][WebWonderings][Musings] [Resources]

So, another busy month gone by.  I'm still in the throws of pre-shipment testing at work but fortunately my code seems to be pretty stable now.  I've even managed to get far enough ahead that I've been able to work at home on other projects while the rest of the gang catches up.  To their credit, their stuff was harder to work on than mine.  I got to design and write my code from scratch.  No legacy code for me.

One of the things I've been working on is getting deeper into the GIMP.  If you haven't seen it yet, take a look at the November Linux Journal.  The first of a 4 part series I wrote on the GIMP was published in that issue.  I also have started to learn much more about how to create interesting effects.  If you have some free time and online-money to burn, check out my GIMP Gallery.  I've put most of my more interesting images there.

This month I'm going to talk a little about how to use the GIMP to create a frame with a 3D metallic or plastic look.  The process turns out to be fairly simply and is one you will probably get to know by heart if you use the GIMP very often.

I'll also cover my notes from SIGGRAPH '97.  I wish I could have done this right after I got back when the conference was still fresh in my mind, but things didn't work out that way.

Finally, I'll be covering JavaScript image rollovers - those cute little image swapping thingies that many people have added to their Web pages.  You can do these using a number of different techniques, but I'm only familiar with JavaScript right now so thats what we'll talk about.

As usual, feel free to send me comments on any of this.

        Disclaimer: Before I get too far into this I should note that any of the news items I post in this section are just that - news. Either I happened to run across them via some mailing list I was on, via some Usenet newsgroup, or via email from someone. I'm not necessarily endorsing these products (some of which may be commercial), I'm just letting you know I'd heard about them in the past month.
 

CurveSoft announces release of FontScope. CurveSoft(tm) is pleased to announce the availability of FontScope(tm): A commercial, high-performance portable rasterizer ibrary for Type 1 scalable fonts sometimes also called PostScript(tm) fonts.  FontScope allows developers and programmers to build Type 1 scalable font support directly into their applications thus roviding a fast, efficient, high quality platform-independent solution to the problem of providing scalable font support.  See http://www.curvesoft.com for further details about the product.  FREE DEMO  You can download a free demo for the Linux(x86) and SunOS(SPARC) platforms from the web page at http://www.curvesoft.com . The demo lets you arbitrarily scale, skew or rotate characters from any Type 1 font you may have and view either the bitmap or the outline on the screen. If you have Multiple Master fonts, it lets you create instances and then view characters from those instances.  If you have access to a Linux(x86) or a SunOS(SPARC) machine, we strongly suggest downloading the demo program since that allows you to test FontScope on your own Type 1 fonts as well as get an idea of thespeed and quality. If unanswered questions still remain, please email them to info@curvesoft.com		The GRASP Environment for Ada 95, C, Java, VHDL GRASP is commercial grade GUI front-end for GNAT, gcc, and JDK. Users can compile, edit, and run programs directly from GRASP. The editor provides complete syntax coloring with user selected colors and fonts.  GRASP provides complete support for generating, editing, and printing Control Structure Diagrams (CSDs) from Ada 95, C, Java, and VHDL source code. The CSD is a diagram intended to increase the comprehension efficiency of source code.  GRASP runs under SunOS/Solaris, IRIX, Linux, and WinNT. It is freely available via the GRASP homepage at  http://www.eng.auburn.edu/grasp       The Texture Library-mirror in Vienna has changed. Please update your bookmarks and links from:  http://klein.htu.tuwien.ac.at/~rs/ to http://freibier.ml.org/texlib/

 

Did You Know?

Q and A

A:  Check out the VRML update on BUILDER.COM:

Q:  Whats the latest news on commercial game development for Linux?

A:  There was an important announcement from Crack.com recently.  Check out

Reader Mail

I read your article in the Linux journal. I was running the older version of gimp (0.5..) and yesterday I wanted to get the new gimp version running.

I configured + compiled + installed.  Everything built fine. Now when I type 'gimp' it says :
gimp: can't load library 'libgtk.so.1'
I have that library built and installed into /usr/local/lib where it is symboliclly linked to 'libgtk.so.1.0.0'. Do you have any idea what I can do to get the Gimp running ? Thanks for your help,

ldconfig generally looks in /usr/local/lib, along with a few other places, for libraries.  To make sure it does, check /etc/ld.so.conf and make sure that directoy is listed there.  Check the man pages for ldconfig if you have any further problems.

If ldconfig -p shows gtk is already installed then the other possibility is to add the following to your .profile:

LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib
export LD_LIBRARY_PATH

kar@webline.dk wrote:

Do you know of a tool that can convert Corel's cdr and wpg formats into a more commonly accessibly format?

It's a little much to buy a package just for that - perhaps you could ask in your column if anybody knows other programs that can?

Tethys <tethys@ml.com> wrote in respose to a Musing last month about NetPBM not having JPEG conversion tools:

Yes, you're right. Netpbm doesn't come with tools to convert to and from JPEG. Maybe it should. However, this is probably because the libjpeg distribution that it would use already comes with cjpeg and djpeg. By default, these already act like ppmtojpg and jpgtoppm. They also accept other formats like GIF, BMP, TGA etc.

My installation diskette for my image256 software was destroyed by coffee, and I've been looking for the software for quite some time now. I have a A4 Tech A4Color AC 4096 hand-held scanner. Do you know of which software can be used to operate the scanner and if possible in Windows environment? Can you tell me where I could download it or get it on the net?

I did a review of all the Linux scanner information I could find in my Graphics Muse column in March of this year.  You can take a look at it starting at the Linux Gazette's (which is the online magazine that carries my Graphics Muse column) table of contents at http://www.ssc.com/lg.

Gregory Gardener <GardenerG@code20nl.npt.nuwc.navy.mil> wrote:

I work for the DOD.  From reading your column in Linux Journal it appears that you have expertise in the area of PC graphics...so I'm directing my question to you. If I wanted to develop a PC software product--an interactive educational system--what PC graphics package is "state of the art" for Linux or Windows?

Don't choose a platform, choose a technology:  Java and/or VRML.  Why?  By doing so you remove the need for your end users to be required to have a particular platform to use your software.  Second, you can more easily fit into any organization no matter what platforms they might already have.

Problems:

1. Java is not quite standard on all platforms yet.
2. VRML is very new and has the traditional problems with new technologies (acceptance, standards, etc).
3. VRML requires the availability of a VRML browser on your end-users platforms.

#3 is a problem.  There are only a few VRML browsers for Linux right now and I don't know much abou them.  Liquid Reality is one - it requires you have the Java runtime environment installed and working.  VRML on non-Linux is actually better supported right now, but thats probably a situation that will be fixed within the next year (again, my guess).

Now, assume you go with these two.  Now you can choose your *own* development platform.  I'd pick Linux.  Why?  Because MS is not being very forthcoming about their support of Java - Linux has embraced it quite happily.  I'm just getting started with Java development so I can't tell you how to use the tools, but the Java Development Kit (JDK), all the way up to version 1.1.3, has been ported to Linux.  There is a commercial Java IDE kit, Vibe, for Linux already.

VRML is like HTML and you can write it using vi or EMACS or whatever text editor you'd like.  Eventually there will be WYSIWYG editors for VRML, but for now it shouldn't be difficult to do by hand.  In fact, many of the current VRML texts are written from the point of view that you are writing it by hand.

Alternatives:  OpenGL/Motif/C.  This is the way I'd do it right now simply because I already now Motif and C and OpenGL is something I've been studying off and on for a few months now.  I really like OpenGL, but if you use Motif/C your application is less portable across platforms (ie its not very easy to port to MS or Macs).  OpenGL has been ported to lots of platforms, but it doesn't provide a windowing toolkit so you have to use it with a toolkit native to your platforms (Motif/X, MS Windows, etc).

Now, this still might not address your problem because "an interactive educational system" can encompass all sorts of display data so its hard to say if Java/VRML will work for you.  What sort of graphics do you need to display?  How do they need to be displayed?  User Interfaces (UI's) is a big area of study that few people hear about because the hype from the technologies you use for them gets more attention.  The problem still exists - which technology do you use?  The answer - whatever one you're most comfortable with for now.  Eventually, you need to get familiar with Java and VRML.

Adding JavaScript rollovers to Web pages.

This month we'll talk about a common trick many people are using in their Web pages today:  image rollovers.  An image rolllover is an image that changes based on user input.  For example, if you look at my GIMP pages (you'll need Netscape 3.x or later for this) you'll see image rollovers in the menus.  When you place the mouse over the text for one of the menu items, the Gallery for example, the text background changes to a light blue or cyan color.  The text is actually an image, not ordinary HTML text.  When you place the mouse over the image it is swapped out for another image, the cyan colored version.  Thats a rollover.

Adding rollovers is actually fairly easy.  First, you need to make two copies of an image (we're taking a very basic approach to this - you can actually create multiple versions of the image to be displayed based on different types of user input).  The first copy is the original, the image to be displayed initially or when no user input is being applied.  The second is the image to be swapped in when the user moves the mouse over the original or clicks on the image.  Creating these images can be done in any number of ways.  Personally, I use the GIMP.  Its well designed for these kind of image processing tasks.

Next you need to understand how JavaScript interprets user input.  These are called events.  There are number of different events that JavaScript can recognize  Not all are applicable to images, however, nor to use with rollovers.  The ones of interest are:
 

Finally, you need to understand how JavaScript references images on a Web page.  The trick to rollovers is to update the correct image - in some cases the image to update is not the image over which the mouse currently rests.  For example, see my contents page where I update a central image when the mouse is placed over images that surround the central image.  That page doesn't do a good job checking which browser is being used and as a result it doesn't work well with non-Netscape 3.x browsers, or even 3.x browsers on non-Unix platforms.  Still, the rollovers do work right on the Linux Netscape browser.

JavaScript references images as objects and as such treats them just like any other object.  Images in a document (documents are HTML pages) are referenced using names, such as
 

document.dog.src

First, lets look at the image's HTML definition:

<A HREF="javascript:void(null)"
onMouseOver="mouseIn('Happy Dog, eh?',); return true;"
onMouseOut="mouseOut(); return true">
<IMG SRC="../gx/hammel/dog-1.jpg" NAME="dog" BORDER=0 HEIGHT=240 WIDTH=320>
</A>

A few other things we should note about this bit of HTML:

1. There are event handlers for when the mouse enters the image area (onMouseOver) and for when the mouse leaves the image area (onMouseOut). We'll discuss the values we associate with each of these in a moment.
2. The event handlers are placed in an anchor with an HREF tag.  The HREF points to a JavaScript function called "void", which is a simple method for doing nothing.  In this case we want only the event handlers to do anything and the anchor to do nothing.
3. We've set the border width to 0 for the image so that no border will be placed around the image.  If we didn't do this a border would be added that used the link and visited-link colors because the image is wrapped in an anchor tag. This way, the image doesn't really look like a link.  Thats good, since its not really a link to anything.

<SCRIPT LANGUAGE="JavaScript">

<!-- Hide script from older browsers

numitems=2;
SampleImages = new Array(numitems);

SampleImages[0] = new Image();
SampleImages[1] = new Image();

// Establish the two image objects we want to use.
SampleImages[0].src="../gx/hammel/dog-1.jpg";
SampleImages[1].src="../gx/hammel/dog-2.jpg";

// Event handler called for onMouseOver event
function mouseIn(input_status)
{
        status = input_status;
        document.dog.src = SampleImages[1].src;
}

// Event handler called for onMouseOut event
function mouseOut()
{
        status="";
        document.dog.src = SampleImages[0].src;
}

// End of hidden script -->

</SCRIPT>

The two event handlers call JavaScript functions.  The mouseIn() function is called to change the image when the mouse enters the image area.  As you see in the HREF tag the event handler associated with this function is the onMouseOver event.  This function has two lines.  The first takes the text passed as an argument and places it on the status bar at the bottom of the browser.  The second line replaces the image referenced as "dog" by the second image in the SampleImages array. Pretty straight forward, don't you think?

The second function, mouseOut(), is called when the mouse is moved out of the image area.  It, too, has two lines.  The first clears the status bar and the second places the image in the first element of the SampleImages array over the image currently in the object called "dog".  In this case, the SampleImages array's first element holds the same image source as the original image displayed in the IMG tag in our HTML source.

A few things to note about the code: I used an array to hold the images I wanted to use for my rollover.  In this simple example I could just have easily used

The image object also contains a number of other properties, such as the BORDER width, image HEIGHT and WIDTH, and the VSPACE attribute for setting the vertical space around the image.  By not specifying the HEIGHT and WIDTH for the document.dog object in the functions I have assumed the new image has the same dimensions as the old.  If this were not the case you might want to update the image dimensions in your functions as well.

Well, that should get you started.  Image rollovers are kinda cool to play with and can add a little pizazz to an otherwise drap Web page, but be careful.  Like most of the neat toys on the Web, too much of anything can be annoying.

   

SIGGRAPH 97 Notes For those that don't know, SIGGRAPH is the yearly conference on computer graphics sponsored by the ACM.  To be exact, SIGGRAPH is the Special Interest Group - Graphics of the ACM and the conference is an extension of that, but in the past few years this conference has grown into a major exposition where all the big names in the computer and entertainment industry show their latest wares.  This year the conference was held in Los Angeles.  It was also my inaguaration into this madhouse.  I should tell you that what follows is not specifically Linux related.  I will try to associated what I saw or learned with Linux as much as possible, but in general Linux is not a powerhouse at SIGGRAPH,  In fact, with the exception of one technical course and a couple of engineers manning booths at the show, I didn't hear anyone talk about Linux.  We have some work ahead of us to get real notice.  For me, SIGGRAPH lasted 4 days, Sunday through Wednesday, even though the conference actually ran for 6 days.  I went as my vacation (sick, isn't it) and after 4 days had been run so ragged I decided my last full day would do something quite and peaceful - so I went to Universal Studios.  SIGGRAPH, for those who've never been, is non-stop energy drain, both physically and mentally.  My first day there I spent going through lines - at the airport, at the car rental (my god, what a mess), and at the conference registration booths.  I had preregistered and so I only needed to pick up my conference materials, but that took about 45 minutes.  Having not been to a conference of any kind since a trip to Interop back in 1991, I was a bit disappointed to see no texts provided for the courses.  In fact, the courses weren't really courses, they were presentations in front of monster sized audiences.  I remember the Interop courses being a bit more personal, more like real classes.  Although I got there late, I did managed to get into the Introductory class on OpenGL that was given by SGI staff members.  It was mostly a discussion on basic OpenGL syntax.  I did manage to wind up sitting right next to Bruce Peren's office mate at Pixar - a very nice fellow named Jim W. (my notes didn't include his last name - hope he doesn't get made at me).  That seemed strange until I found out that Pixar sends just about everyone to SIGGRAPH.  A few days later I saw Larry Gritz but didn't get a chance to talk to him.  Like most people at SIGGRAPH, he was in a hurry to get somewhere.  The rest of the day I spent wondering around the LA convention center looking for where other classes and events would be held.  There was a career center downstairs where people could post resumes and companies posted job positions.  Xi Graphics had a posting there, but other than that I didn't see any Linux-related postings.  By the time I got back to my hotel it was about 9pm - and I had yet to eat that day.  One word of warning to future attendees:  take snacks!  Finding food can be hard and what you do find can be awful (at least in LA by the convention center) and you will need the enegy!  Being a running, I should have brought a few energy bars with me.  Well, thats a lesson learned.  The next day I went to the Advanced OpenGL course, also taught by SGI staff.  This was a more interesting class to me, since it talked about techniques a bit more than just syntax.  There were some good discussions on using environment maps (mapping an image to surface), projecting textures, and billboarding.  This last one is a cool technique I wasn't aware of.  An image is mapped onto a planar surface and that surface is programmed to always face towards the viewer.  It is sort of like taking a photo of a tree, cutting away all of the photo except the tree, and pasting it to a popsicle stick.  The stick is then rotated so that the photo of the tree always faces you.  It simulates a 3D effect without all the overhead of huge polygon counts for a real tree.  This technique works well for scene objects that are in the distance but looses its effect up close.  You can see billboarding effects in many games.  Another effect they covered was the creation of fire.  They used a noise function along with what are called "quad mappings" to blur regions of the noise.  They then added some turbulence and an absolute value  function to create veins in the fire.  It turns out you can do something similar with the GIMP by using the solid noise filter, followed by some blurring and waves adjustments, and cleaned up with a combination of curves/color balance adjustments (for color), and sharpening (for details).  When I got home I went to visit my sister and her kids and watched them play Super Mario Bros on their ... Sega?  Anyway, having just returned from SIGGRAPH I could see the effects they talked about in the advanced OpenGL course.  It was very enlightening.  One of my reasons for going to the OpenGL courses was to get an idea for how multimedia applications could be built for Linux.  An alternative to an OpenGL based multimedia kit would be one based on VRML. I attended the VRML Demo SIG on Monday night and it had some very cool demos using VRML 2.0.  If based on CD-ROM this could be a very good cross platform environment for multimedia packages (edutainment, for example).  The current drawback is that the demos required 2.0 capable browsers, but there should any number of those showing up in the next year or so.  According to CGW (freebie that was dropped at my hotel room - very nice touch), Netscape 4.0 has the SGI Cosmo Player embedded in it now.  I haven't been able to get this to work yet, however.  Since I have a 2.0 kernel installed now I can try some of the other VRML browsers available.  Maybe I'll get to that next month.  -Top of next column-		More Musings...   GIMP Tip - 3D metal and plastic borders.
		On Tuesday I arrived 1/2 hour early to take the 3D Animation Workshop.  I was about 20th in line of 100 or so, but somehow managed to miss getting a little blue ticket.  Despite the people at the door having seen me there for 30 minutes waiting (before others had showed up) I still didn't get in.  I was very pissed.  This class was one of my prime reasons for my going to SIGGRAPH.  Hint:  if you get in line at SIGGRAPH, make sure you check with the people in front of you to see if you need to have some special form, ticket, or sacrificial offering to get in.  My fallback course was the GUI/Multimedia class.  There really wasn't much multimedia discussion - mostly it covered how to design GUI intefaces.  Lots of design criteria which I'd learned by fire long ago.  When they finally did get around to multimedia they talked about the lack of standards and how most applications are reinventing GUI rules.  This was my assessment before going to the class, but it was nice to have it reinforced by someone who studies such things for a living.  Next I skipped out to Mark Kilgard's OpenGL w/Windowing systems class.  Pretty good stuff.  Brian Paul talked about MesaGL a little as well as some portability issues.  Both are very nice guys - I chatted with them a bit after the class.  One thing Mark mentioned was that there arent' any toolkits specifically built with OpenGL, which makes sense.  There is an overhead in creating widgets using OpenGL that would make it unsuitable for most applications.  Instead, OpenGL is integrated into applications using some other toolkit, like Motif, XForms, or Tk, using a toolkit specific OpenGL-capable widget.  Later on I stumbled into the Java3D class for a few minutes.  Looks interesting but I didn't get too many details.  Sun is saying a beta release of a sample implementation of the specification  is due out in December of 1997 but the course speakers said they hoped to have it out much sooner.  They also expect that since the specification is public and being openly developed that sample implementations from non-Sun sources are likely to show up before the Sun release.  Or at least they expect that - they didn't have any information on any specific implementations.  At this point I haven't heard of any Java3D announcements from my usual sources.  I talked to a couple of recruiters at the career fair, mostly just to get a feel for what companies are looking for these days.  They mentioned that the best way people can avoid getting lost in the crowd of candidates at the fair is to submit resumes based on job postings from the corporate web sites.  If you go next year looking for work, check out some of the participating companies job listings before heading to the career fair.  First day of the show for me was Wednesday.  It looked more like a cross between a theme park ride and a disco, especially when Intel rolled out their MMX "clean-room dancers".  SGI offered a 1/2 hour demo ride with their O2 systems.  Cute demo but really didn't give a good feel for what you were really buying.  After talking to some of the show staff and taking the demo ride I still don't know exactly what applications come with the stock O2 box.  The $4995 price offered to conference participants wasn't bad, however.  Certainly compares well with a comparably equipped PC running NT.  There is a freeware/shareward CD available for O2 developers according to one of the engineers I talked to, but I didn't see it.  The SGI O2 demo showed me one thing which I had hoped I'd find out from the show:  how to create multimedia applications.  In fact, after the demo a White Paper was passed out explaining how the demo was put together.  Some items, like the networking that allowed real-time texture mapping of live video, are not really possible on Linux boxes with available software.  But most of the rest is:  OpenGL rendered in Motif widgets was used for the GUI and OpenGL and VRML were used for interactive 3D displays.  OpenGL is available commercially and with MesaGL.  Motif is available commercially. I'm not certain VRML is available for integration into applications via an API, but animation could be done using MPEG.  There are some MPEG API's, such as MpegTV's API.  ImageMagick's PlugIn distribution appears to have an MPEG library, but I don't know if it provides a decent enough API for commercial multimedia applications.  After seeing the SGI demo and the VRML SIG demos I've changed my mind about VRML - its ready for prime time.  I wouldn't recommend it for network applications yet, at least not Internet based applications, but for private networks (such as kiosks) or CD-based multimedia applications it could offer some unique possibilities.  Wacom was at the show and a few of the staff didn't get defensive when I brought up Linux (it all depends on how non-abrasively you bring up the subject). They were even quite apologetic about not having any info on which tablets were supported by XFree86.  They now have a pad that is actually a full-color flat screen - you simply draw on it like you do with the tablets.  Very cool, but currently runs > $2200.  Cosmo Software (new division of SGI) staff were pleasant as well, mentioning that it might be possible to provide someone with the source to port the Cosmo Player to Linux.  The player (which plays VRML 2.0 scenes) is free for various other platforms.  I guess they just need someone to ask for it and provide some credentials to prove they could do a decent port.  Unfortunately, I'm not certain who to talk to about this.  Huge crowds formed around the Apple booth right at the entrance.  One demo I caught caused an OS lockup, but for the most part people were quite excited about the products.  I don't know much about Mac's so didn't stay long.  However, like LInux, Mac is an underdog in the OS world and I can't help but hope they survive.  The one thing about this trip that stood out for me was LA itself.  It was Hot.  Outside.  Inside.  In classes.  In the car.  In the hotel.  Everywhere.  Hot. LA also has an ugly downtown compared to others I've seen.  It also apparently has very few restaurants anywhere near the Convention Center.  A 6 block walk in Colorado is no big deal, but a 6 block walk in LA in August....*sigh*.  Food was a big problem for me.  I hope the same problem doesn't exist in Orlando next year.
submitted resumes based on job postings from their web sites.

 
  The following links are just starting points for finding more information about computer graphics and multimedia in general for Linux systems. If you have some application specific information for me, I'll add them to my other pages or you can contact the maintainer of some other web site. I'll consider adding other general references here, but application or site specific information needs to go into one of the following general references and not listed here.
 

Linux Graphics mini-Howto  Unix Graphics Utilities  Linux Multimedia Page  Some of the Mailing Lists and Newsgroups I keep an eye on and where I get much of the information in this column:  The Gimp User and Gimp Developer Mailing Lists.  The IRTC-L discussion list  comp.graphics.rendering.raytracing  comp.graphics.rendering.renderman  comp.graphics.api.opengl  comp.os.linux.announce

Future Directions

Let me know what you'd like to hear about!