System Design
Resume:
Keith Packard
abp8gn@r.postjobfree.com
Experience
HP (via Compaq) (12/01-present)
Member of the Cambridge Research Laboratory. Research projects focused on user
interfaces in all guises, from tiny embedded computing devices running either custom
operating systems or uCLinux through multi-machine projector-based media walls.
X window system research included the design and implementation of the X Com-
posite and X Damage extensions which provide for manipulation of application presen-
tation through external applications enabling a wide range of new user environments.
This novel new architecture provides capabilities far in excess of those available in
other window systems while remaining compatible with existing applications.
Designed and implemented the cairo 2D rendering library supporting X, PostScript,
PDF, Windows and Mac OS X output while providing best in class rendering using
a PDF 1.4 compatible imaging model. Encouraged adoption of this library by many
major free software projects, including Gtk+, Mozilla, Mono and Inkscape.
Developed new font naming and selecting library called fontcon g to unify appli-
cation use of fonts in free software systems. Promoted library to free software projects
by providing relevant patches to their code, including Mozilla, Qt, Gtk+, Tk and X.
Fontcon g is now used across all of these applications as the fundemental font con g-
uration mechanism, eliminating a common source of font access dif culties.
SuSE, Inc (11/99-11/01)
Enhanced the XFree86 implementation of the X Window System. Designed and im-
plemented a new rendering extension. Built a new X driver architecture for embedded
devices. Wrote and presented numerous papers at technical and industry conferences.
Worked with other hardware and software vendors on X related projects, including the
Linux port for the Compaq iPAQ.
Developed new text rendering mechanism to improve readability on LCD screens
by modulating intensity of individual color elements. Integrated this algorithm into X
applications to provide best in class text presentation for window system graphics.
Network Computing Devices, (3/95-10/99)
Designed and implemented X-based products including X terminals, Windows-based
X servers and X-based multi-user NT systems. Interacted directly with NCD customers
1
and resellers to promote X based solutions in general and NCD products in particular.
Presented several technical papers and tutorials.
MIT X Consortium (3/88-5/92)
Member of a small group (2-7) of people directed by Robert Schei er responsible for
the development and standardization of the X Window System. Involved in almost
all of the related standards efforts, both within the X Consortium and with national
standards bodies (ANSI, IEEE). In charge of X server development at MIT for release
3, 4 and 5. Presented many technical papers and tutorials at an international collection
of X-related conferences.
Tektronix Inc. (12/83-3/88)
Designed early X terminals. Worked with a team developing an X10R4-based inte-
grated C development environment based on an incremental C compiler
Education
Reed College, Portland, Oregon. BA, Mathematics, 1986.
Publication
All of these are available from http://keithp.com/ keithp/talks.
Twin: An Even Smaller Window System For Even Smaller Devices
With embedded systems gaining high resolution displays and powerful CPUs the de-
sire for sophisticated graphical user interfaces can be realized in even the smallest
of systems. While the CPU power available for a given power budget has increased
dramatically, these tiny systems remain severely memory constrained. This unique en-
vironment presents interesting challenges in graphical system design and implementa-
tion. To explore this particular space, a new window system, Twin, has been developed.
Using ideas from modern window systems in larger environments, Twin offers overlap-
ping translucent windows, anti-aliased graphics and scalable fonts in a total memory
budget of 100KB.
Ottawa Linux Symposium, July 2005.
Getting X Off The Hardware
The X window system is generally implemented by directly inserting hardware ma-
nipulation code into the X server. Mode selection and 2D acceleration code are often
executed in user mode and directly communicate with the hardware. The current archi-
tecture provides for separate 2D and 3D acceleration code, with the 2D code executed
within the X server and the 3D code directly executed by the application, partially
in user space and partially in the kernel. Video mode selection remains within the X
2
server, creating an arti cial dependency for 3D graphics on the correct operation of
the window system. This paper lays out an alternative structure for X within the Linux
environment where the responsibility for acceleration lies entirely within the existing
3D user/kernel library, the mode selection is delegated to an external library and the X
server becomes a simple application layered on top of both of these. Various technical
issues related to this architecture along with a discussion of input device handling will
be discussed.
Ottawa Linux Symposium, July 2004.
The (Re)Architecture of the X Window System, with Jim Gettys
The X Window System, Version 11, is the standard window system on Linux and UNIX
systems. X11, designed in 1987, was state of the art at that time. From its inception,
X has been a network transparent window system in which X client applications can
run on any machine in a network using an X server running on any display. While there
have been some signi cant extensions to X over its history (e.g. OpenGL support), X s
design lay fallow over much of the 1990 s. With the increasing interest in open source
systems, it was no longer suf cient for modern applications and a signi cant overhaul
is now well underway. This paper describes revisions to the architecture of the window
system used in a growing fraction of desktops and embedded systems
Ottawa Linux Symposium, July 2004.
Xr: Cross-device Rendering for Vector Graphics, with Carl Worth
Xr provides a vector-based rendering API with output support for the X Window Sys-
tem and local image buffers. PostScript and PDF le output is planned. Xr is designed
to produce identical output on all output media while taking advantage of display hard-
ware acceleration through the X Render Extension.
(This is the rst paper on the cairo graphics library, which was originally called Xr
for X Rendering library).
Ottawa Linux Symposium, July 2003.
X Window System Network Performance, with Jim Gettys
Performance was an important issue in the development of X from the initial protocol
design and continues to be important in modern application and extension develop-
ment. That X is network transparent allows us to analyze the behavior of X from a
perspective seldom possible in most systems. We passively monitor network packet
ow to measure X application and server performance. The network simulation en-
vironment, the data capture tool and data analysis tools will be presented. Data from
this analysis are used to show the performance impact of the Render extension, the
limitations of the LBX extension and help identify speci c application and toolkit per-
formance problems. We believe this analysis technique can be usefully applied to other
network protocols.
Usenix Annual Conference, San Antonio Texas, June 2003.
3
Font Con guration and Customization for Open Source Systems
Font con guration and customization has traditionally been left to each application.
Fontcon g is a library designed to provide a common system that can serve to ease
application development and provide users with the ability to con dently install new
fonts with the expectation that they will be used by most applications. Fontcon g pro-
vides the ability for multiple con guration interfaces to affect a wide range of systems
without requiring custom code for each new system. Fontcon g provides a range of
services to allow applications to pick those appropriate without being forced to use the
entire interface. Wide acceptance of the Fontcon g mechanisms will improve system
consistency without requiring a radical redesign of existing applications.
Gnome Users and Developers European Conference, Seville, April 2002.
The Xft Font Library: Architecture and Users Guide
The Xft library was written to provide X applications a convenient interface to the
FreeType font rasterizer and the Render extension. As FreeType provides for no con-
guration or customization, Xft also performs this task. Xft provides new font naming
conventions, sophisticated font matching and selection mechanisms and suf cient ab-
stractions to permit common applications to bene t from Render extension based text
output while still working on X servers without support for this extension.
XFree86 Technical Conference, Oakland, November 2001.
Design and Implementation of the X Rendering Extension
The X Rendering Extension addresses many of the short-comings inherent in the core
X rendering architecture without adding signi cantly to the protocol interpretation or
implementation burden within the server. By borrowing fundamental image composit-
ing notions from the Plan 9 window system and providing sophisticated and extensible
font rendering, XFree86 is now much more able to support existing applications while
encouraging new developments in user interfaces. More work remains to be done in ar-
eas where best practice is less well established, including precise polygon rasterization
and image transformations.
Usenix Annual Technical Conference, Boston, June 2001.
The X Resize and Rotate Extension - RandR, with Jim Gettys
The Resize and Rotate extension (RandR) is a very small set of client and server ex-
tensions designed to allow clients to modify the size, accelerated visuals and rotation
of an X screen. RandR also has provisions for informing clients when screens have
been resized or rotated and it allows clients to discover which visuals have hardware
acceleration available.
Usenix Annual Technical Conference, Boston, June 2001.
Translucent Windows in X
The X Translucent Window Extension is described which solves the general translu-
cency problem by assigning alpha values for pixels in occluding windows. These values
4
are used to blend the occluding window contents with the occluded region for display.
The details of managing translucent window hierarchies, re-parenting translucent win-
dows and X visual differences between blended pixels are discussed.
Atlanta Linux Showcase, October 2000.
Ef ciently Scheduling X Clients
The X server is charged with providing window system services to many applica-
tions simultaneously, and needs a scheduling mechanism to distribute it s limited re-
sources among these applications. The original scheduling mechanism was simplistic
and caused graphics-intensive applications to starve interactive applications.
A new scheduling mechanism has been designed which fairly distributes time among
the requesting applications while at the same time increasing performance by a small
amount. Descriptions of the original and new scheduling mechanism and empirical
measurements demonstrating the effects of scheduling within the X server are included
along with a discussion on how the design was arrived at.
Usenix Annual Technical Conference, San Diego, June 2000.
A New Rendering Model for X
The rise of inexpensive Unix desktop systems in the last couple of years has led to the
development of new user-interface libraries, which are not well served by the existing
X rendering model. A new 2D rendering model is being developed to serve this new
community of applications. The problem space and proposed solutions are discussed.
Usenix Annual Technical Conference, San Diego, June 2000.
Font Support in WinCenter Pro: Creating an Application Speci c
Font Server.
One of the challenges in implementing WinCenter Pro (an X based multi-user NT
system) was to ef ciently render text. Using strike-format bitmaps or scan converting
each glyph into a list of rectangles consumes signi cant network bandwidth and takes
signi cant time for the X server to render. For reasonable performance the system
must use the X text rendering primitives. Substituting an existing X font for each NT
font fails to preserve application appearance. WinCenter Pro exports the NT fonts to
the X server using the X Font Service Protocol providing both high performance and
pixel-perfect results. The overall architecture along with some of the technical issues
involved are presented.
Tenth Annual X Technical Conference. The X Resource, O Reilly & Associates,
Issue Seventeen, 1996.
A Pseudo-Root Extension: X Window System Nesting on a Budget.
The notion of encapsulating a screen inside a sub-window has been around a long time.
Rob Pike s layers went as far as possible; the only way to create nested windows was
to run a new copy of the window system from within a window. Release 6 provided an
5
X server which could do the same (Xnest). However, both of these suffer from perfor-
mance problems as each client request must be delivered over two network connections
and be interpreted by two window system servers. The X Window System provides for
nested windows already, and has only a few references in the protocol to which win-
dow is the magic root of the window hierarchy on the screen. By changing which
window appears to be the root in the context of a particular client, a full-speed window
system encapsulation has been achieved. Functionality to encapsulate additional global
resources is also included.
Ninth Annual X Technical Conference. The X Resource, O Reilly & Associates,
Issue Thirteen, 1995.
Design and Implementation of LBX: An Experiment Based Stan-
dard
Unlike other X Standards efforts, the design of Low Bandwidth X (LBX) could not be
done in a vacuum. With the goal of LBX to provide a usable X environment via an ex-
tremely limited bandwidth channel, the only way to effectively design the system was
to experiment with different ideas and measure which actually worked better. Some of
these experiments could be done with pencil and paper, other required extensive devel-
opment. Many design decisions were made using parts of the eventual implementation.
The process, along with the experiments, results and design will be presented together.
Eighth Annual X Technical Conference. The X Resource, O Reilly & Associates,
Issue Nine, 1994.
The Layout Widget: A TeX style Constraint Widget Class
The X Toolkit geometry management process is extremely exible and powerful; how-
ever the existing composite widget classes make it dif cult for the application devel-
oper both to simply design an application layout, and even more important, to make
the layout work in a wide variety of environments.
The Layout widget class is described which uses a stretch/shrink model similar to
TeX to constrain the layout of an application in a manner which allows the geometry
of the children to respect the desires of the application designer, while adapting to its
environment, both in terms of the changing geometry allocated to the widget, and to
the changing needs of the child widgets. The speci cation of the child layout is entirely
contained in a resource which is interpreted at run time.
Seventh Annual X Technical Conference. The X Resource, O Reilly & Associates,
Issue Five, 1993.
Using XTrap to Help People with Manual Disabilities
The XTrap extension provides a mechanism to interpose a complicated application
between the X input devices (pointer and keyboard) and the X server. Using this mech-
anism, an interesting system for reducing the amount of manual ability required to
operate X clients is investigated and compared with other systems.
Sixth Annual X Technical Conference. The X Resource, O Reilly & Associates,
Issue One, 1992.
6
X Selection Mechanism
While the existence of the selection mechanism in X may be wide known, the details
of using it are not. This paper, while not a research paper per se, attempts to join
the technical details with some practical experience. Along with this paper, a sample
application was written which provides a working example of the ideas presented here.
Fourth Annual X Technical Conference, Boston, 1990.
7
Contact this candidate