UK Web Archive Technology Watch

Opera 10 released

Opera have released a new version of their web browser: Opera 10.

Opera have released detailed changelogs for each platform that supports the browser. Identifying those changes which are most relevant for preservation purposes is not straightforward - to some extent, that depends on what your preservation objectives are. Some of more technical changes for the Windows release include: (text taken directly from the changelog; not independently tested):

'Opera Presto 2.2 rendering engine:
The Opera Presto rendering engine has been updated to version 2.2 further enhancing its reputation for stability and security. It is identified by the following User Agent ID string located at Help > About Opera > Browser identification: Opera/9.80 (Windows NT 5.1; U; en) Presto/2.2.15 Version/10.00

Pretty-printing of unstyled XML:
Opera 10 incorporates "pretty-print XML", which is now the default styling for unstyled XML and uses the unstyledxml.css style sheet in the Styles sub-directory of the Opera installation directory.

Acid3 test:
The Acid3 test is the third in a series of test pages written to help browser vendors ensure proper support for web standards in their products. Acid3 is primarily testing specifications for “Web 2.0″ dynamic Web applications. It also includes some visual rendering tests, including webfonts.
Opera 10 scores 100/100, pixel perfect on the Acid3 compliance test.

Web specifications support:
CSS (Cascading Style Sheets)

• Support for the CSS3 color:transparent value has been improved.
• Full RGBA and HSLA opacity support is now included for an easier way to make Web page features transparent. This is accomplished by the addition of a fourth argument to HSL and RGB, namely alpha transparency and results in RGBA and HSLA values. See this Opera reference.
• Support has been added for the CSS Selectors API. This feature makes the selection of DOM elements a lot simpler. It includes partial support for namespace resolver features which allow you to work with mixed namespace documents and select elements based on their namespace; see this Opera reference.
• CSS files must now be served with the correct MIME type ("text/css") in Strict mode, or they will be ignored.

HTML 5

• HTML 5 algorithms have been implemented for detecting charsets in HTML.
• HTML 5 end-tag and start-tag parsing, whitespace parsing, and DOCTYPE parsing have been added.
• HTML 5 <canvas> elements can now export images to the JPEG format.
• HTML 5 <canvas> transforms are applied when building a path, not when painting it.

SVG (Scalable Vector Graphics)

• FPS (Frames Per Second) support: Opera enables you to manipulate the speed (frames per second) of your SVG animations using JavaScript by supporting the SVGElement.currentFps and SVGElement.targetFps properties. These properties respectively read and control SVG frames per second.
• SVG fonts in HTML support: Use SVG font files to style your text using CSS (in both HTML and SVG files).

Web fonts

• font specification (@font-face CSS at-rules) with the font-family descriptor
• src descriptor with local and remote
• TrueType (TTF)
• OpenType (OTF)
• Scalable Vector Graphics (SVG) font support'

There are also several changes that directly impact on the user experience, including an inline spell checker, Opera Turbo - which increases your internet bandwidth speed on slow connections using data and image compression technologies, new skins, and a change to the visual tab settings so that tab thumbnails can be previewed.

Windows System requirements:
These requirements only apply to Opera. Third-party plug-ins such as Flash, Java, and others may have additional requirements.

Recommended configuration for Opera:

• Windows XP or later
• 256 MB of RAM
• 100 MB of free disk space

Minimum configuration for Opera:

• Windows 2000 on a Pentium II
• 128 MB of RAM
• 20 MB of free disk space

System files used by Opera for Windows are listed here, alongside an extensive set of support documentation.

Linux system requirements:
'The Standard C++ Library is required to run Opera for Linux. Most Linux distributions have it available. On Debian or Debian-based distributions you can install it by typing something like apt-get install libstdc++6. You can also get this library from most Linux vendors, or you can download and compile it from gcc.gnu.org/libstdc++/. You also need to have an X11 implementation installed.'

Mac system requirements:
OS X Panther (10.3) or higher on an Intel- or PowerPC-based system.

Plug-ins:
Opera supports most Netscape Compatible Plug-ins. An overview of the main plug-ins is provided on the Opera plug-ins webpage

(Note: Opera 10 is not to be confused with Opera Unite, covered in an earlier post).

Mac OS X Snow Leopard

Mac OS X is the operating system for Apple Macintosh computers. On August 29th, Apple released version 10.6, otherwise known as Snow Leopard.

Apple have been producing Macs with Intel processors since 2006, having previously utilised Power PC processors. Snow Leopard does not support the Power PC architecture. Organisations opting for hardware emulation as a preservation solution will need to bear this in mind if they wish to run Snow Leopard. PowerPC Macs include Power Macs, PowerBooks, iBooks, some iMacs, and all eMacs.

Installation requires 1GB of memory and 5GB of available disc space. Most of the built-in applications in Snow Leopard are designed to utilise a 64-bit architecture, though they will run in 32-bit mode on machines with 32-bit processors (such as IA-32 Intel processors). Snow Leopard does not support writing to HFS (Hierarchical File System) partitions, though it does enable read-only access.

The Apple website identifies a number of ‘incompatible software’ applications that SL will automatically (during installation) move to an ‘incompatible software; folder at the root level of the hard drive. These are:

• Parallels desktop v2.5 and earlier (Parallels)
• VirusScan v8.6 (McAfee)
• Norton Antivirus v11.0 (Norton)
• Internet Cleanup 5 v5.0.4 (Smith Micro)
• Application Enhancer v2.0.1 and earlier (Unsanity)
• AT&T Laptop Connected Card v1.0.4, 1.0.5, 1.1.0 (AT&T)
• Launch2net v2.13.0 (NovaMedia)
• iWOW plug-in for iTunes v2.0 (SRS)
• Missing Sync for Palm Sony CLIE Driver v6.0.4 (Mark Space)
• TonePort UX8 Driver v4.1.0 (Line 6)
• ioHD Driver v6.0.3 (AJA)
• Silicon Image SiI3132 Drivers v1.5.16.0 (Silicon Image)

The website also identifies a number of other software applications that are restricted from opening as they are known to quit unexpectedly:

• Parallels Desktop v3.0 (Parallels)
• VirusBarrier X4 v10.4.4 and earlier (Intego)
• SPSS 17 v17.1 (SPSS)
• Director MX 2004 v10.2 (Adobe)
• EyeTV v3.0.0 to 3.1.0 (Elgato)
• Ratatouille v1.1 (THQ)
• AirPort Admin Utility for Graphite and Snow v4.2.5 (Apple).

In such cases, Snow Leopard prevents the application from opening and presents a dialogue box to notify users of incompatibility issues. However, several of these applications are regarded as ‘old versions’ – for example, Parallel’s Desktop application is now in version 4.0 – and most of the current versions of the applications appear to work okay, particularly as more time passes and developers are able to produce newer, compatible versions.

FITS-ing tools to support preservation

Harvard University Library has released FITS, the File Information Tool Set.

FITS combines the strengths of a number of different open source file identification, validation, and metadata extraction tools by normalising, comparing, and consolidating their output into a single XML output file. Tools currently supported by FITS include:

• JHOVE - the JSTOR/Harvard Object Validation Environment
• Exiftool - a platform independent library and application for reading writing and editing metadata in image and audio/visual files
• NLNZ Metadata Extractor - a preservation metadata extraction tool for various different file formats
• DROID - a file format identification tool
• Ffident - a file format identification tool
• File Utility  - a file format identification tool

Other tools are under evaluation for incorporation into FITS in the future. When tools produce conflicting output (for example by producing non-identical file format identifications), FITS can resolve this by normalising output and referring back to previously entered data such as user preferences and a configurable ‘format tree’.

FITS has been designed for easy integration into preservation workflows. It is open source, written in Java, is compatible with Java 1.5 or higher, and it extensible to enable the addition of new tools in the future. Tools added to FITS are wrapped in XML so that the FITS program can invoke them, capture the output, and normalise it. 

Forwards planning with Firefox 3.5

On Tuesday June 30th, Mozilla launched Firefox 3.5. Firefox is the second most widely used web browser (the first being Internet Explorer), currently occupying over 20% of the market share. Firefox is a free, open source web browser managed by the Mozilla Corporation. It was first released as v1.0 in November 2004; v2.0 was released in October 2006, and v3.0 in June 2008.

V3.5 adds a variety of new features and support for other technologies, most notably support for the video and audio tags defined in the forthcoming HTML 5 specification, which will enable videos to be played without dependence on plugins. Very few webpages currently contain such coded audio and video elements, though we can expect it to increase when HTML 5 is approved.

Firefox 3.5 also supports:
-    location aware browsing: if activated by the user, the browser uses the PC’s IP address in conjunction with Google Location Services to deliver location-specific information
-    private browsing mode
-    tab restore mode – closed tabs can be reopened at the same page within a single session

This version uses a new Javascript engine named Tracemonkey. Firefox 3.5 is reputed to be twice as fast as v3.0 and ten times as fast as v2.0.

Commentators have noted that the clock is now ticking for Firefox v3.0 as Mozilla’s policy is to support older editions only for six months after the release of a new version. Security patches and support for Firefox 3.0 will therefore cease at the end of 2009.

Opera Unite: DIY hosting?

Opera Software has launched a new version of its Opera browser called Opera Unite. Currently available as an alpha build, Opera Unite offers users the ability to connect with each other in a more direct manner, rather than through third party servers. Opera Unite browsers have an integrated server component, allowing for services to be hosted directly from a user’s local computer rather than run by third parties. Services currently include website hosting, file sharing, a notes facility and a chatroom.

 

Opera Unite can be run on all platforms that support the Opera browser, including Windows, Unix and Mac. Hosts must have the Opera browser installed, but visitors can access the services offered by an Opera Unite installation via any modern browser. Anyone can make Opera Unite services, but they will only become ‘official’ OU services once reviewed and accepted by Opera Software to ensure they comply with Opera T&C and do not contain any errors.

 

We will be discussing the impact that hosting content on local PCs rather than third party servers may have on our web archiving processes in due course.  

DROID 4.0 - a quiet release?

DROID 4.0 has been released by The National Archives.

DROID performs automated batch identification of file formats and their versions, using internal and external signatures. It does this using information stored in the PRONOM technical registry. PRONOM is regularly updated with new signature information and DROID can be configured to automatically download updated signature files when they become available.

DROID is a platform-independent Java application that comes with a documented and public API to facilitate integration with other systems. The release of DROID 4.0 appears to be a quiet one and so far there is little accompanying documentation. Development has been carried out in conjunction with the EU-funded PLANETS project

The news item on the sourceforge website notes the following changes in v4.0: 'The focus of the latest major release is the inclusion of the DCS (Digital Continuity Service) and Planets, Collection Profiler. DROID now runs in two modes. The original file identification mode, and the new “profile” mode that allows users to obtain file format information gathered from large distributed sources of digital files, providing users with aggregated statistical data and reports to help them take appropriate management decisions regarding risk associated with such large collections of files.'

Web Curator Tool v1.4.1 released

A new version of the Web Curator Tool (v1.4.1) was released on May 27th 2009.

The Web Curator Tool (WCT) is a tool to help manage the selective web archiving process. It is a collaborative development between the British Library and the National Library of New Zealand, initiated by the International Internet Preservation Consortium.

Version 1.4.1 of the Web Curator Tool is now available from Sourceforge. This version includes many bugfixes and an upgrade to Heritrix 1.14.1. There are also some new features and performance improvements that were commisioned by The British Library and developed over two iterations during September-October of 2008 and February-March of 2009 by software developers from Oakleigh Consulting in the UK.

This version is primarily a maintenance version of the tool. It includes nine enhancements to the tool, and sixteen bug fixes. Enhancements include:
-    Allowing indexing of 'stuck' harvest results to be restarted;
-    Upgrading to Heritrix version 1.14.1;
-    Improving the speed at which Target records are saved.

Note that Heritrix 1.14.1 was a bugfix release with a small number of enhancements, including writing WARCs with the declared format version "WARC/0.18".

A full WCT release history can be found on the Sourceforge website.

JHOVE 1.3 available

JHOVE has been updated with improved identification and validation support for PDF, TIFF, and XML files.

JHOVE (the JSTOR/Harvard Object Validation Environment) is open source software capable of identifying, validating, and characterising digital objects. Accurate file format information is vital for the long term preservation of digital objects. The functions supported by JHOVE can be useful at various stages across the life cycle of digital objects, particularly in early ingest and in part-validation of subsequent preservation activity.

JHOVE is a Java application that should be usable on any Windows, Linux, or OS X platform with an appropriate J2SE installation. This portable, platform independent design is also conducive to long term preservation. It has an extensible plugin architecture, with every file format having its own module. It offers the following standard modules:

• AIFF-hul: Audio Interchange File Format
• ASCII-hul: ASCII-encoded text
• BYTESTREAM: Arbitrary bytestreams (always well-formed and valid)
• GIF-hul: Graphics Exchange Format (GIF)
• HTML-hul: Hypertext Markup Language (HTML)
• JPEG-hul: Joint Photographic Experts Group (JPEG) raster images
• JPEG2000-hul: JPEG 2000
• PDF-hul: Page Description Format (PDF)
• TIFF-hul: Tagged Image File Format (TIFF) raster images
• UTF8-hul: UTF-8 encoded text
• WAVE: Audio for Windows
• XML-hul: Extensible Markup Language (XML)

JHOVE examines the internal structure of a file to identify its file format and version, rather than relying on the file format extension.

Changes to JHOVE in v1.3 that affect the PDF, TIFF, and XML modules are detailed in the JHOVE v1.3 release notes.  JHOVE v1.3 can be downloaded from the Sourceforge website.

[Release first announced on the File Formats Blog - thanks Gary]

Chrome goes 2.0

On May 21st 2009, a stable version of Google Chrome 2.0 was released. 

Google Chrome v2.0 (or, strictly speaking, v2.0.172.28) was first released as a beta in March 2009 (as v2.0.169.1). The Google blog identifies a small number of enhancements, including:

• An improved New Tab page – when new tabs are opened in Chrome, they are populated with thumbnails of most frequently visited pages. Users now have the option of deleting thumbnails from the tab, an option that was not previously possible;
• A full screen option – the browser interface can be hidden when viewing some types of material, such as videos or presentations;
• Form autofill – shows information previously entered into similar forms to make online forms quicker to complete;
• Increased stability – over 300 bugs have been fixed since the initial launch of Google chrome in September 2008;
• Increased speed – Chrome v2.0 uses a new version of its Javascript engine, V8, so interactive pages run significantly faster than in previous releases. 

Chrome 2.0 (both beta and stable releases) uses a new version of the WebKit open source rendering engine, v530.8. Webkit is a powerful application framework that also provides the rendering engine in the Konquerer and Safari browers.   

Chrome 2.0 does not provide support for third party extensions, though it does have embedded support for  plugins such as the Adobe Flash Player.

Chrome 2.0 runs on MS Windows XP SP2 and later. Linux and Mac versions are in development.

What's Up, Flock?

Flock v2.5 was released on May 18th 2009. Flock is a web browser tailored to facilitate use of social networking sites, particularly (but not exclusively) Twitter, Facebook, Bebo and Flickr. The Flock blog states:

‘ Flock 2.5 makes sharing and discovering content fast and easy. You just drag and drop URLs, photos, videos, text or other things you find on the web to a friend’s Twitter, Facebook, MySpace or other profile in Flock’s People Sidebar and it’s instantly shared. And now Flock comes with Twitter Search right in MyWorld, so you can keep up to date on all the topics you’re most interested in and save them in the best place possible, your Flock browser.

Last but not least is FlockCast, an effortless way to share information across your social networks. With FlockCast, you can broadcast anything from blog posts and picture uploads to Tweets and MySpace status updates directly to your Facebook profile. And when you share a URL in a Twitter message, Flock automatically shortens the URL.’

Flock is based on the Mozilla Firefox browser. Its architecture is pictured (literally) at the Flock developer wiki, where the Flock source code is also available.

Numerous issues remain to be discussed about the benefits of using this browser in a web archive. The current attraction of the browser stems from its integrated functionality to facilitate contribution and sharing of content across the web by users. In contrast, it is conceivable that most typical users of the web archive (as it currently stands) will be more focused on experiencing content, rather than adding it. Though addition of content (eg through annotations) is not ruled out, it is likely to draw on different functionality. The actual Flock user experience may also be of interest to future users of the web archive. Comments on these scenarios are of course, very welcome.