400 likes | 545 Views
Implementing Web Standards across the institution: Trials and tribulations of a redesign Patrick H. Lauke, Web Editor, University of Salford. Institutional Web Management Workshop / Birmingham - July 2004. Workshop programme. Workshop aims. At the end of the session participants will:
E N D
Implementing Web Standards across the institution:Trials and tribulations of a redesignPatrick H. Lauke, Web Editor, University of Salford Institutional Web Management Workshop / Birmingham - July 2004
Workshop aims At the end of the session participants will: • Be familiar with “web standards" • Have gained an insight into the advantages of “web standards” • Be aware of potential problems, and approaches to resolve them
So why am I here? • Web Editor for University of Salford • Small central team, 30+ devolved web authors • September 2003 University relaunched new “web standards” based core site • A few trials and tribulations along the way • Many web people considering move to web standards • Here to share my experiences • Not a guru, don’t have all the answers – simply a method that worked for us • Hoping to generate good discussion
Setting the scene: what do we mean by “web standards” Technical: • working to a common, agreed syntax (W3C spec) • no proprietary markup - compatibility • generating code that validates (so you can have your little badge on the site?) “Ethos”: • Return to basic principles: HTML for content, CSS for presentation • semantic/structural markup (no validator for that!)
Case study: trials and tribulations of a redesign – the Salford experience “How we got from there…
Case study: trials and tribulations of a redesign – the Salford experience (cont.) … to here”
Case study: background • University website redesigned December 2000 • first effort by External Relations to bring consistent look and feel • external design company • happy to say: I didn't do it! (started in January 2001)
Case study: problems with the site Purely from design point of view: • Compliant with CI, but tied to print campaign • Dominant design feature in its own right • “Naff”/”Kitsch”/{insert expletive here} • Structurally confusing: “where am I?”
Case study: problems with the site (cont.) Technical issues: • Cluttered code: FONT, TABLE • HTML not made for “round corners” = more markup to fake it • As result: templates cumbersome
Case study: problems with the site (cont.) • Pages didn’t print well • Need for “printer friendly” versions
Case study: problems with the site (cont.) …and many more problems: • graphical buttons • dropdown navigation(accessibility and “spiders”?) In short: a mess. But…we’ll keep it for a while. Fixed some issues, but most problems remained…
Case study: fast forward 2 years… • Beginning of 2003 University started CI review • Tightening of lax guidelines, creation of new ones • Web would need “face lift” • Stricter rules for Faculties/Schools/etc: adopt the templates! Do you: • Simply slap new facade on decrepit old building? • Make a fresh start, learning from past mistakes?
Case study: why “web standards”? • Nowadays: “web standards” buzzword • At the time: just trying to follow best practices • Separation of content/presentation • Lighter code – quicker download times • Accessibility concerns (SENDA/DDA): making site work in maximum number of browsers – no proprietary markup • What about next redesign? • “work smarter” / “web design on a shoestring”
Case study: why XHTML specifically? • Separation of content/presentation can be achieved with HTML4.01 just the same • Requires “personal” discipline • Stricter syntax for XHTML removes most/all presentational markup - validation brings more things to light • Future plans of CMS – repurposing content: XHTML is XML, so simple tools available (XSLT)
Case study: why abandon tables? • Syntax of XHTML still allows tables (rightly so) • “Ethos” however: tables for tabular data, not layout • Using pure CSS driven layout: increased flexibility for future redesigns • Same page / different delivery channels (screen, print, etc)
Case study: approach - structure “tabula rasa” – start from scratch • New development server • Inventory of current content • Working out new structure, discarding old/redundant content • Initially, simply copied pages to new directory structure
Case study: approach - template Ideal situation: • Create page structure • Style the structure
Case study: approach – template page structure • Concentrated on identifying “functional blocks” • Branding (logo) • Search box • Navigation • Breadcrumb trail • Content • Footer • Tempting, but don’t think about what it looks like!(however, think about order in which blocks appear in code) • Directly translates to XHTML – DIVs (or appropriate block level elements – FORM) • Try choosing most “semantically appropriate” elements (e.g. navigation as list) • Validate
Case study: approach – template style • Creating stylesheet probably took longest • Ideally, XHTML “frozen” • However, occasional need to revisit XHTML: re-ordering elements, adding “hooks” for specific styling
Case study: approach – template style (cont.) • Develop for most compliant, then work backwards • From general to specific (e.g. rough block position, before tackling padding/margin) • Validate What about old browsers?
Case study: approach – populating the template Now bringing it all together: • Content from existing site extracted from pages(sounds easier than it is: find/replace, retagging, etc) • Same process: • Create most appropriate XHTML • Where necessary: new page/section specific styles • In theory: simply “pop it into the template” (plus few manual tweaks)
Case study: approach – populating the template (cont.) • “Relatively easy” to create beautiful CSS driven layouts with known, “frozen” content(cfr. CSSZenGarden) • Real-world content offers “interesting” challenges • Often requires revisiting content XHTML, or even template XHTML/CSS
Case study: approach – let’s get dynamic • Static pages converted, but not forgetting database driven areas (e.g. news/events, course finder) • Mostly simply updating server-side scripts’ output • Databases containing badly formed HTML: • UPDATEing db tables after cleanup • Solving problem at the root: ensuring HTML data well formed (if not valid) before committing to database: Editize and “sanity checks”
Case study: launch • After final validation and browser testing: launched September 2003 • Set up redirections / rewrite rules on server for new structure • Monitoring error logs / 404s
Case study: does the design solve original problems? Design: • In line with tighter CI • More neutral: allows page-specific design elements • Feedback: “professional” / ”polished” • Less confusing for visitor (breadcrumbs, visible navigation)
Case study: does the design solve original problems? (cont.) Technical: • Separation content/presentation • “lighter” code (20%-30% saving or better) • Templates far easier
Case study: does the design solve original problems? (cont.) • No need for “printer friendly” pages (print stylesheet)
Case study: does the design solve original problems? (cont.) • No need for graphical buttons • Navigation now pure list of links: accessible, “spiderable”
Case study: problems experienced • Majority due to inexperience with XHTML/CSS – learning by doing • Choosing semantically most appropriate elements not straightforward (but XHTML is flawed!) • Authoring tools still not good enough: DW code view, glorified text editor with FTP client • Flaky CSS support and browser bugs: most annoying • Testing on multiple platforms not always possible: Mac and different versions of IE
Case study: what would I have done differently? • Learning XHTML/CSS while going along resulted in frequent re-starts (now would probably take less time) • Not using XHTML 1.0 Transitional, but straight to Strict • Not gone far enough in terms of “semantics” • Although minimal use of “modularisation” (includes), would go further: more includes, template engine (SMARTY)?
That was easy… …now for the hard part!
Hard part: getting web authors to follow • Redesign of core site was fairly easy: single developer • How to get 30+ web authors, with varying skill levels, to follow my lead? Answers on a postcard…but in the meantime, this is the approach we’re taking…
Hard part: approach • All sub-sites physically hosted on same server • Created templates, based closely on core site templates • Use of global includes for header • Stick: new web publishing guidelines, stricter rules (plus teeth to enforce them) and best practice recommendations • Carrot: all imperative guidelines taken care of automatically if web authors use templates
Hard part: approach (cont.) • Education, education, education: replace generic “how to use Dreamweaver” with tailored staff dev sessions • Web strategy: ensuring Faculties/Schools/etc recognise technical requirements of post, and resource accordingly (still growing teeth to enforce) • Any 3rd party supplier needs to adhere to standards as fundamental requirement Majority of sub-sites now transitioned to new design, however this is not the end…
Hard part: continuous QA • “But it was valid when I first created it…” • Validation of XHTML/CSS as routine, second nature • Making it as simple as possible: URI based validation, using right tools for the job • Automatic checks (based on server access logs) and alerts (e.g. “validator to RSS”) • Any “external” data sources either fixed at source, or run through filters (TIDY)
Conclusion Brian Kelly: “People may be interested to know how you managed to get your homepage to validate as XHTML 1.0 Strict” Hmmm…through hard work. • No magic bullet • steep initial learning curve • “Paradigm shift” • Continuous monitoring / QA
Doing it for yourselves: exercise • Split into groups • Identify problems of implementing web standards in your own institutions • Discuss solutions/strategies to overcome them • Feed back
Contact Patrick H. Lauke Web Editor Marketing & Communications External Relations Division University of Salford E-mail: p.h.lauke@salford.ac.uk Web: http://www.salford.ac.uk Personal site (on web standards, css, experimental techniques,etc): http://www.splintered.co.uk