Social Weave

Email forms a primary social network according to Jeremiah Owyang. Since more functionality seems to be needed, custom scripts would be appropriate. Asked LinkedIn about practices and it echoed warnings of impending Google Wave. Upcoming open-source cross-browser extensions and API will take advantage of built-in media and contact handling, e.g. using HTML5. It is reportedly coded using Google Web Toolkit (GWT), which is also discussed in  Hanson and Tacy 2007.

Other wiki additions included Croll 2009 on performance, Hurley 2007 on security.

Can use HTTrack to copy web site content. Parts of Test Info Space can be downloaded into a project for off-line review.

A web service can be run directly from a device using Opera Unite for browser-based collaboration which includes darknets.

Twitter beached the fail whale and graduated from recent celebrity status to geopolitical journalism concerning Iran election. There may be more exceptions in the midst.

Basalist

Consider whether an ideal tester can emerge from a distributed repository such as Test Info Space, or what a utility of this nature would be like. The effort might lead to a knowledge base to support development and validation efforts, if not as straightforward as a spell-checker for an editor, then at least to assist in browsing, documentation or presentation. A private cloud or internal network might add or replace web content.

Assume that the purpose of the tool is for debug, test and diagnostics. It would have some form of search, e.g. keyword or semantic, a Q&A ability or directory, recommendation generator, a logging mechanism for tracing or reuse, and scripting for automation. It is a composite of sources including history, state of the art, and research. Users can lookup concepts to find descriptions and supporting links, or test cases to match solutions by example. It can provide comparisons by similarity, counter, or uniqueness. This might be done at a high level of methodology or fine-tuned by fields, domains, characteristics, indications, and tradeoffs, or deduced from issues such as data, analysis or symptoms. Rankings can be sorted by title, author, editions, style, level of difficulty, citations, prerequisites, mathematics, and so on.

The present wiki includes pages for each title showing authors, publication date, category, tags, links to content, external links, possibly a book site, other media, and user comments. Sources usually have some combination of table of contents, glossary, index, notes, hyperlinks, bibliography, code and exercises. These may be in any of various formats, e.g. PDF, HTML, hardcover, paperback, flash or e-reader. External pages or bookmarks may provide a booklist linked to the wiki pages. There may also exist other reviews or a synopsis. Authors have biographies, publications, blogs, organizations, teams, social networks and feeds. Tools have sources, possibly open-source, vendors, costs, and dependencies or chains. Projects can be active or archive.

New wiki pages include Guernsey 2009 on databases, Ogrinz 2009 on enterprise mashups, Resig 2006 on javascript, and Newkirk 2004 on test-driven development (TDD).

Image: TIS topic list

Sharper IM Age

How do we figure out where to look for answers amid the offerings from computation, search, wikis, social networks, semantic web, and so on? Datasets differ. Would be nice to handle user preferences, history, context, word-based problems, voice, and local state of knowledge. Caches allow them to provide up-to-the-minute FAQs. Timestamps are able show historical answers. Text-driven user input can be cut, piped to functions, and the answer returned or posted somewhere, e.g. an editor. Questions can be sent to all and merged or integrated, or the APIs can make them recombinant. The hardware scales from cells to chips. Cloud-based services seek to be more platform than killer app because various industries each have idiosyncracies, so there will likely be multiple solutions. Also started a basic wolframalpha experiment as way to observe learning curves for utility and user. There seem to be more variables than controls initially, so will need to characterize what is (un)settling and how.

Human-based textual image parsing this week included Farrell-Vinay 2008 on test management, McDonald 2008 on QA, Burns 2007 on security, Holzmann 2003 on SPIN model checker, Krepinevich 2009 on futurism, and Oreilly 2009 on the microblog.

“When you have eliminated the impossible, then whatever remains, however improbable, must be the truth.”

Warewiki

Tools evolve and drive inventors to new efforts. The wiki is probably no exception. It can be used to comment software and track bugs. Static content pages can be translated to wiki format using automated keyword hyperlinking and tagging. It seems to be the case that, by adding a few simple additional field types, they can be made more active. For example, a selector entry might be switched between values to use different other pages. So we enter the world of wiki-based computing. Documentation applications handle system descriptions, designs, metadata and statistics. It is not impossible for them to add operation and debugging. Each page is a memory snapshot or state of a set of values which can include hyperlinks and tags. A series of these can be used to drive activity just as a set of source code modules can be an instance of an algorithm, or machine components can be an engine. These modules are versioned and there can be a selector field parameter to string the correct choices together. This gives some flexibility in either pre-specifying order of execution, or allowing a test-driven approach to exercise combinations until a result is obtained. Automated editing or refactoring is more complex; this can be content-driven or result-driven. It may be possible for the wiki to edit itself to adapt to some external context or usage. A processing medium would have the ability to execute from wiki description or code to binary. This implies a ware descriptor language. Applications might be imagined for emulation, virtualization or hypervisors. This allows digital experiments to compare implementations prior to production. A wiki can also then become an operational component of a system. Pieces may be upgraded to other media or machines. Consider the wiki OS, model, processor, grid, and infrastructure. These are based on a built-in CMS, alerts, activities, artifacts, and abstractions. Analyzer output can generate wiki pages which may also be used as input to resume from. Dimensions of page relationships include hyperlinks, timestamp, writer, as well as content matches under search patterns, and other rules. Set the clock and see what establishes domains for the entry title.

Recent reading list Barrett 2008 on mediawiki, Ford 2008 enterprise s/w, Souders 2007 performance, Hamill 2004 unit testing, Crispin 2002 extreme programming, and Peck 2009 on Gimp.

New socnet test activity started on Daney’s SaRnet.

Image: MediaWiki

Instant Wiki

wiki

Tiddlywiki has an html file that can be installed on local storage as a wiki app. The snapshot shown above is from a laptop instance; another was created on a flash drive which allows portability between devices. Google chrome was not able to save edits. Flock seemed to work okay.

Subtle UIs

There is a common logic behind various website designs such as Google, Amazon and Imdb. The goals are to achieve clarity and make the user’s life easier. Navigation and search are key. They reinforce the user’s ability to quckly get the idea and complete the task. The site ID, home page, search, sections and utilities are easily found. This is the thesis of Don’t Make Me Think, Krug, 2005, a popular book about empirical web usability testing. Wikis are another aspect of this quickness approach. Also see author’s site and additional resources.

Feed Frame

In previous posts, we have looked at feeds as updates for resources and commentary. It can also be useful to convert their current contents into a web page. For example, links can be inserted into a wiki by pasting from rss2html sourced from a bookmark site such as delicious using selected tags. This is the case for The Art of Debugging with GDB, DDD and Eclipse, Matloff and Salzman, 2008 based on gdb tag, and Lab Notebook from notebook tag. The wiki itself supplies another feed. It is possible then to translate between content management system (CMS) types such as bookmarks, wiki, database, spreadsheet, documentation or website.

Edited smartpen wiki for activities.

Read JUnit in Action, Massol and Husted, 2003. Wiki.

“Game Testing All in One”, Schultz & others, 2005, 511pp describes how software testing techniques can be applied to games. After a humorous introduction orients the reader to the differences between gamer and tester approaches (e.g. rules 1 & 2 taken from Hitchhiker’s Guide, & X-Files or Ice-Nine, respectively), the authors reveal their knowledge of some popular titles as well as methods for improving the user’s experience. Book chapters have exercises answered in appendices. The contents demonstrate how to dissect a game for testing, and drive the software and hardware components in both well-defined and ad-hoc fashions. Testers get to know the games inside out and often spend weeks working through the various phases until release. Beta, online and multiplayer testing are also covered. Automation tools are used (since the game AI usually isn’t responsible for the testing part, too). Document, form and test flow diagram templates are detailed in an appendix. There are links to some commercial and public-domain tools to support the effort, e.g. documentation in a wiki.

Testing of mobile devices is covered in a recent webcast. Equipment costs can be lowered using combined analyzer and generator components local to a test station or connected using web-enabled LXI and IEEE 1588 timing.

Human Research Engines

Reviewed Information Trapping: Real-Time Research on the Web, Calishain, 2006, 328pp. Information traps allow the user to set and forget search details until new results are ready for evaluation. Datastreams may be sourced from search engines, news, blogs, web pages, tagged content, site links, multimedia or conversation groups. Domains may contain commercial or government sites, for example. The input can be filtered through email, client-based, or web-based tools. Results may then be edited and published to email alerts or mailing lists, documents, web sites, or cell phones. Advanced search queries using keywords are covered in depth for engines such as google, yahoo, ask or msn live. These usually involve nouns to monitor data about a thing, person or place. A couple of chapters discuss various aspects of RSS feeds. The traps are refined over time for variables such as what is of interest, where to look, how often, and in what format to review results. Many links are given to web resources in each category. Author’s blog. This covers techniques for automating search using existing utilities. It complements some of the other references discussed previously for structuring a system to emphasize findability, creating scripts to analyze information, or mashups.

Another aspect that has become popular is social-powered search. This involves people who locate or rank results before they are presented to the user. Chacha, Mahalo and recently Searchamigo implement this. Kindle NowNow does it for a portable book reader. Younanimous Aftervote and URL do meta-search. The latter and Sproose allow users to vote on preferences. Mosio handles mobile phone texting. 50matches limits to sites selected by users. myshoppingpal is recent and specialized in its subject matter. Goodsearch donates to user-preferred charities.