“Linux Device Drivers, 3rd Edition”, Corbet, Hroah-Hartman & Rubini, 2005, 636pp, outlines how software used to interface machines and devices is constructed. Drivers interface application software with the hardware through kernel functionality. The release 2.6 device model supports power management, user space communication, hotplugging, device classes and object lifecycles. The authors talk about how these are built, handle memory mapping and DMA, do signaling such as interrupts, run in a device bus and multi-process environment, and are refined. Chapters have a Quick Reference summary at end of each which lists header files and C function calls. Sample driver code is given for block, character and network devices represented by sbull, scull and snull, respectively. Details such as firmware download and portability issues are discussed. Particular hardware dependencies are listed across various processors, e.g. for I/O ports. Emphasis is on most popular drivers such as PCI, USB and, to a lesser extent, ISA. Kernel debugging talks about options to enable debugging modules, printing methods (printk, console, klog, syslog, /proc/kmsg), querying (ioctrl, /proc files), watching (strace, gdb, kgdb), system faults, hangs (sysrq key, kernel profiling), the user-mode linux port (UML) which runs a kernel in a user-space process, and linux trace toolkit (LTT) which allows event tracing, dynamic probes (DProbes). See commons.

Media and network devices that exist in multi-operating system environments, or require a particular OS, may also be of interest. The Ubuntu version of Linux 2.6 can be run on Windows machines booted either from a partition, CD, or file. Wubi is a downloadable install that brings up a desktop. Lubi can convert it to a partition. Contents written to the wubi filesystem are visible to windows, after reboot, and vice-versa, or can be shared via a removable device such as USB flash, network connection, server or web site.

Tag Teems

In “Tagging: People-powered Metadata for the Social Web”, Smith, 2007, 216 pp, metadata is stored in fields of larger records. These might appear as web site interfaces, but also be used for custom utilities that can analyze contents based on the metadata values, or schemas which combine reserved types and unrestricted metadata labels. Tagging allows users to label and sort their web and computer data items and to look at others’ labels or items for which there are common labels. This book looks at the phenomenon, how it originated and has been implemented, and some particular sites that organize social bookmarks, media sharing and personal information. The author discusses javascript frameworks, and shows examples of sql code that can be used to improve on the performance of server-side scripts. There are links to tag sites, browser extensions that group items across sites, and APIs for user applications. Timed tags are shown in Viddler. Open-source tools include FreeTag.

Tumble logs, or tlogs, are a quick way to link a series of data objects of various types into a weblog. See Tumblepedia. The Tumblr API shows how these can be read or written from other applications.

Looked at “The Linux Programmer’s Toolbox”, Fusco, 2007, 656pp. This discusses some ways to measure use of memory, virtual memory, paging statistics, inter-process communication (IPC), multi-processing, and to do performance profiling. There are some examples of shell scripts. It also offers an overview of Linux tools for configuration, development, debugging and tuning. There are links to open-source download sites. It discusses the philosophy, derived from Unix, of combining a lot of special-purpose tools to accomplish more general-purpose functionality. There are techniques that can be used within programs themselves, or via the shell commands. The Linux architecture that makes this possible is reviewed. Alternate ways of doing some of the more complex things are shown with their relative merits, e.g. getting details about opened files such as their inodes.

Importals

It was the case earlier this month that a popular web URL bookmarking site was not in service for a couple of days starting during the weekend. The site eventually returned with links intact, but the unexpectedness highlighted the need for backups and alternatives. Have therefore tried export/import features across several sites for redundancy. An HTML file can be saved from the user browser favorites or a site. A tag is similar to a folder on the desktop and can have multiple values on the web. Links can usually be assigned public or private. Tag density may show up in the list or cloud.

Mozilla Firefox can save from the bookmarks, organize bookmarks, file, export option to an HTML file, as can IE from file, import and export. Each can import from its own file. Firefox 2.0.0.11 can import IE 7.0.5730.11 file and vice-versa.

CiteULike and Connotea link academic papers and sources. Was able to import into the latter using HTML and it had quarantine period where links are kept private. Blogmarks had public registration off so did not run the test.

Blinklist can save to a local HTML as well as other formats. The user page can be public and the cloud can be embedded in other sites. The options, import links feature can get bookmarks from del.icio.us or furl. This site can import a Firefox browser file, and the tag is derived from the bookmark folder name. Firefox can import the blinklist file and descriptions are derived from notes. IE could not read the blinklist file.

Del.icio.us imported HTML favorites okay. Resultant links were private.

Furl imported HTML after a retry. Entries were then public. It shows tags in a “my groups” list with an entry count for each.

Netvouz imported HTML, public links were same. The tag “web_2.0″ became separated into “web” and “2.0″ so had to edit those entries to “web2.0″. (Netvouz itself could not be bookmarked into BlinkList at that time.) This site offers a direct import from del.icio.us after login.

Simpy imported browser file HTML okay, but it did not have the tags. It has a separte notes organizer. This site can import directly from del.icio.us after login.

These can also be tried using Opera or Safari browsers.

Rawsugar did not seem to have import/export capabilities.

“Embedded Linux Primer: A Practical, Real-World Approach“, Hallinan, 2006, 576 pp, talks about developing software to run some common devices. This book is an introduction to the use of Linux for embedded device operation and software development. It has tutorials for installation, debugging, and porting. Enhancements for kernel version 2.6 are outlined. Real-time issues are highlighted, e.g. preemption and profiling. Discusses memory technology devices (MTD) subsystem and journaling flash file system (JFFS). Appendices have more details about U-Boot and BusyBox commands, SDRAM interfacing, and BDI-2000 hardware JTAG debugger. Has pointers to many open-source sites including some emerging standards. Newer types of memories and file systems might require additional sources. Test equipment is often a form of embedded device which adds applications for measurement, storage, comparison and display, for example. These can also be networked for monitoring and alerts.

Reviewed “Visualizing Data”, Fry, 2007, 382pp. This book presents a seven-step process to ensure understandable information display: acquire, parse, filter, mine, represent, refine, interact. The approach is of interest to those who need to convert their frequent or complex information needs into executable code. Demonstrates the Processing front-end API for web site analysis and discusses the software design issues and approaches. Starts with simple introduction to chart drawing, curve fitting, maps, trees, and file handling techniques. Advances to applications that parse text files to show lexical maps, and display user activity across web site pages. Code can be ported from, or interface between, java and Processing or vice-versa. Techniques are shown for scraping data, downloading files, outputting to documents, handling binary streams, or using site APIs. For example, shows source behind displaying a graph of labels, in different colors and sizes, that might represent a census distribution or bookmark tag cloud. Many links refer to compatible free or product software including spreadsheets, databases and network analyzers. Some of this also applies to dataflow for inter-machine services. Author site: www.benfry.com. Demonstration video. Processing site.

Reviewed “Critical Testing Processes: Plan, Prepare, Perform, Perfect”, Black, 2003, 608 pp.