Change log

JavaScript menu tree has been externalized. Previously JavaScript used to produce navigation menu tree items was included in the actual page, it has now been moved to external JavaScript file required some additional logic to js itself). This makes handling changes more straight forward, simplifies the xhtml page code and also decreases the file size. In the name of accessibility I also added fallback mechanism to produce simplified navigation menu tree for browsers with no JavaScript support (or JavaScript support turned off).

Well, this seriously sucks. Turns out Firefox is happy to devour xmp-files using responseXML (XmlHttpRequest) from the local drive but doesn't want to process the very same files from the web server. After loads of trial and error, I found out that overrideMimeType to text/xml is the rescue.

I finally took the time to write a Ajax-code necessary to pull random quote from an xml-file. I've been doing the very same for years using Perl-script, so the reason for all the trouble was to get my head around on how to use XMLHttpRequest, a cornerstone of Ajax. This isn't used at6 the moment though.

As a side product I started to ponder whether I could use the same technique to randomly pick an image. Well, to pick up just the image would be easy (doesn't really differ from randomly picking up the quote). What differs though, that in order to properly handle the image, I would need to have access image metadata as well (for caption) and size attributes. Given that I had earlier employed Exiftool to automatically generate xmp-sidecar files, I have that data available in xml-files. So basically what I needed to do was to

1. Select random item from the list containing a list of files I want to rotate
2. Request the xmp-file of the chosen image and extract the metadata
3. Return properly formatted html to be written into specified div within the calling document

Sounds simple enough. And surprisingly, it didn't turn out to be much more complicated that it sounded. Well, in principle anyway. Once I figured out how to traverse the xml-structure returned as XMLHttpResponse, pulling off the functionality I was after in a separate proof-of-concept project was easier than anticipated. Xmp-files contain several namespaces, something that I fully expected to cause major headaches. Turns out they don't. Ok, so far so good.

My next step of action was trying to implement that in real-life scenario. Which is about the time when I ran into some peculiar functionality. Turns out I had tumbled on notorious domain restrictions of Ajax. Well, sort of anyway. Everything is part of the same domain, but trying to read from file using relative addressing so that I would have to go one directory level backwards just didn't seem to work. Which is a major PITA as I pretty much would have to do exactly that. After considerable amount of frustration and wasted time, I found out that this issue doesn't really exist; it is only encountered when not running the site on a web server.

That being said, I don't particularly like to run a local web server just for this very purpose. Luckily, it isn't too difficult to circumvent (well, few things are if you happen to know how to). In this particular case I found that in Firefox, you can bypass strict local file security policy by going into Firefox config panel (type 'about:config' to address bar) and set property "security.fileuri.strict_origin_policy" to 'false' by double clicking on it. That allows you to continue to access local files that are 'outside' of the directory structure you loaded from via XHR. IE seems to have the same limitation, not sure whether it can be circumvented somehow (and if it can, how). Neither Opera or Chrome seem to have such a restriction. Anyway, now that this is sorted, I'll probably tidy up the code, then implement it so that I can rotate images on mountain page.

When tinkering with that, I found out that this site causes quite a few CSS errors, which clutters browsers error console, thus making it more difficult to track down real issues. At least the majority, if not all, CSS issues are caused by hacks used to overcome and circumvent browser-related bugs. It is very likely, that I will get rid of those hacks in the near-future; if someone still used either IE 5.5 or old Netscape browsers, they are clearly gluttons for punishment anyway. The same probably goes for conditional comments used to target various older IE versions, they make the code cluttered and difficult to edit with some editors. Anyway, I moved some dirty IE hacks from the main CSS to IE targeted CSS-file, which minimizes the implications of the issue.

So much for my namespaces not being a problem. It appears that you can make namespaced files to work either in

• Firefox and IE using getElementByTagName("namespace:element"), which really isn't correct according to spec
• Firefox, Opera and Chrome (probably loads of others) but not in IE by using getElementByTagNameNS("namespace", "element"), which is correct according the spec but does not work in the most common browser
• Cross-browser by applying either browser-aware processing in JS, some sort of hacks or implement external wrapper component (such as Sarissa) that make it work cross browser.

Obviously any one of these options is poor.

I finally got my around on how to retrieve details from external xml-file using ajax instead of cgi. I might very well use this approach later on to replace old cgi script currently used to pull random quote from text file. The solution seems to work in Firefox, IE and Chrome.

However, probably the more important feature this opens up is that I think I have what I need to rotate images randomly. Well, just to rotate images randomly would be far simpler, but this approach give me access to xmp sidecar-files which is necessary for me to be able to use details from image metadata stored in aforementioned sidecar files without the need for any sort of help from server-side software. There's still quite a bit of things I need to solve before that is possible though (most importantly having to figure out how to use different namespaces and traverse rather complicated xml-structure with DOM and JavaScript).

Several improvements in layout and functionality

• Nicer looking and more compact layout of mountain details.
• Fixed and simplified several processing functions in publish scripts, especially related to images.
• Changes publish scripts to use relative paths and inline tools (Saxon-B and Exiftools) to make whole site including publishing stuff self contained.
• Added some new links
• Addedc new images to mountain pages and replaced some other images with new ones. Extended the usage of xmp metadata.

Several improvements in layout and functionality

• Nicer looking and more compact layout of route details.
• Automatically generated events table on mountain level.
• Adedd Mountain Project to context menu
• Added links to Google Maps for peaks with gps details.
• Added display of route type

I finally got around of implementing support for image metadata. My solution uses ExifTool to extract embedded metadata from images and storing it into xmp-sidecar files. This is done as a batch process using Windows batch file. After that publishing xsl pulls the details from xmp-files to be shown in the site.

Lots of updates in Bernina page.

I went on and replaced the "web 2.0" links available on some of the pages with JavaScript menu, which I reckon is less obtrusive and ultimately better from the maintenance point of view as well, as it would not require changes in markup. The script used on the site is based on JavaScript Context Menu by Luke Breuer. The sample has been changed quite a bit though. Unfortunately my tweaking seemed to have broken compatibility with Internet Explorer 7, though. I tested the script with Firefox (3.0.7 and 3.1 beta), Internet Explorer 8 and Safari 4 beta. The best of this approach is that whenever the integrated services syntax requires changing or if I want to add new services, all I need to do is change the JavaScript. At this point, the integrated services are:

• Google Search
• Wikipedia
• SummitPost
• Flickr
• Picasa
• Google Maps

Links are now currently automatically generated on mountain list. This requires help from Windows Powershell script as xslt is not capable of processing file tree. Powershell script basically gathers the list of files contained within a single directory and stores the result as export-clixml -format directly available through Windows Powershell. The rest is handled with xslt that reads a name of a peak of the table, then finds out in which document that peaks exists and generates a link to it.

I've also been playing with the the idea of replacing the "web 2.0" links with JavaScript menu, which I reckon would be less obtrusive and ultimately better from the maintenance point of view as well, as it would not require changes in markup. Given that the markup is automatically generated at the publishing time, this isn't very important at all though. I have a working solution taking use of RightContext script of Harel Malka. However, I would like to trick that up a bit to get rid of some attributes which currently need to be in markup (these would be pretty trivial to generate automatically in the script in my use case).

Finally, there's been several additions and corrections on many mountain pages.

I added plenty of new functionality for creation of mountain and route lists. These can be used to gather data from mountain page to generate automatically list of peaks located on a specific area to main article page grouped by location (sub-area) with automatically generated links for ease on navigation. All this is is achieved through rather complicated xsl(t) and xpath trickery. Xslt 2.0 just kicks some serious but.

Books now link to Google Book Search.

Route and mountain lists now pull information (such as location, grading and effort) from mountain page. Of this info, currently only grading info is used. Site now support adding crumb path to headers (currently only used on North America -page). At the same time, the generation of those lists is now more sophisticated in linking, it now only links the mountains and/or routes if the mountain/route actually exists on this site. If not, plain text is shown. This addition also had some impact on layout.

I went on and added more web 2.0 functionality in a form of "mashed-up search". Basically this adds capability to launch direct context-sensitive searches to Flick, Picasa, Wikipedia or SummitPost. This feature is currently only used on North America page.

Changed Asia, British Isles and South-America pages to pure xml structure. In a process, plenty of typos were corrected and some additional details were provided, especially for Cordillera Blanca and Patagonia sections. To automate the process of (xhtml) to xml transformation I had to employ Regular Expresion and following-siblings functions of xslt 2.0 to cover my earlier poor design choices. Allso, all book references on those pages are now pointing back to Bookcatr database export file, thus reusing references to same books is vastly easier to achieve.

Rather complete overhaul of current news page. Added plenty of details which previously only existed in font page left side column to current news as page as well. Also removed all dead links and added some new articles.

Since this site is hosted by a web host not providing me access to imageMagick or any other similar tools, I have no way to read metadata directly from the image files to be used on website. For quite a while I've been searching for a solution making it possible for me to read the metadata embedded into files in my photo cataloguing apps rather than having to enter the details outside of the files. One might guess that since photo management gods at Adobe are using their xmp extensively in their tools such as Lightroom, Photoshop and Bridge, there would be host of apps able to write all metadata to sidecar xmp files. One would be wrong. Well, there's no real shortage of apps capable of outputting xmp sidecar files for Raw images. But when you need an app capable of writing xmp sidecar files for jpg files, the pickings are significantly slimmer. Granted, you are supposed to write xmp info into jpeg itself, but having that info there is not a great help when trying to output that info into web page. Some apps are capable of writing xmp sidecars for single jpeg files, but try to do that for several images all at once and your options are very limited indeed. Luckily, ExifTool with ExifTool GUI is up to the task at hand. Seems that I no more have an excuse for not implementing this.

• ExifTool GUI

I caught a nasty flu, therefore have has not much to do during last couple of days. So I revamped some of the publishing xsl by consolidating the code and fixed some areas I had previously broken. A side result of this is also some additional features. I also implemented xml schema to my external link file to prevent me from doing typos etc. in element and field names. Probably the biggest change is that I finally got rid of using definition list in link lists, mainly because support for hReview and xFolk require wrapper element around link content which seriously breaks IE. And strictly speaking it wouldn't be valid markup either. So off with dl, in with ul's. Byprocuct of this is more semantic markup as well. This prompted some fiddling with layout css's as well.

Added links to articles previously featured at front page only to matching sections of the site as well.

Added new tags to index page news. Moved some of the inline links on main page to external links repository file. Added some details to some links, most importantly element date on links. Added support for hReview in both (event list) paragraph content (full text review) and link list (linked review).

I recently noticed that Internet Explorer (including Internet Explorer 7) did a pathetic job of displaying this very site; (at least) all unordered (ul) and ordered lists (ol) were displayed incorrectly. This seems to be caused by IE:s inability to process such elements correctly whenever they are located within floated elements. Which is rather sad given that many pure-CSS layouts rely on floats to build the layout. Which is the case with this site as well. To make matter worse, there's no real solution to remedy this. Fortunately upcoming Internet Explorer 8 (currently available as beta 2) seems to finally fix this.

However, since lists are heavily used on this site and lack of bullets and inproper indents can seriously impair the readability of some of the pages, I added IE conditional comments along with CSS targeted to IE7 that fix this problem. At least up to the point where the layout is at least pretty close to what it should look like.

• About Conditional Comments
• Conditional Comments
• Internet Explorer conditional comments
• Internet Explorer conditional comments
• Targeting IE Using Conditional Comments and Just One Stylesheet

I also changed the mime-type to application/xhtml+xml which is what is recommended for xhtml 1.1. I am well aware that this may cause issues with old browsers. Furthermore, references to xhtml 1.1 schema are now added to html-root element.

Two new articles on index page.

Corrected few broken links. Removed pages that have been replaced by new ones.

Added two new items to front page. Adedd few new links to front page.

Added some new links to front page. Added support for XFN microformat. Fixed links and some other data in several mountain pages.

Added new photo gallery. Changes photo gallery tags to be read from external tag repository file (the same one used by front page news). Moved inline links from photo gallery page to external links repository page.

Moved quick links from front page sidebar to external links repository file. Added tags to several links stored in links repository file. Simplified the structure of links repository file.

Added front page news. Changed favicon to different style. Fixed issues of front page (3-col layout) media="print" css so that printing is again available.

Added favicon

Plenty of added and corrected details on Mont Blanc area.

I decided it was time to move away from my old and trusted work-horse, namely xslt 1.0, and replace it with it's more powerfull big brother, xslt 2.0. Granted, the support for xslt 2.0 is seriously lacking, but since I am only using xslt on the publishing side, that doesn't really matter. Therefore, off with msxml and in with Saxon B. To let xslt 2.0 flex its serious processing muscle, I also implemented loads of internal linking, automated link generation and grouping. Pennine Alps page is the only place where new and improved xslt stylesheets are in use at the moment, however they replace all xslt 1.0 stylesheets shortly.

Change of xml parser also prompted me to change the publishing procedure, formerly I used Windows batch file and msxsl command line tool to perform necessary transformations, with Saxon the principle is the same, but Saxon being Java-application, some changes are necessary to publishing scripts as well.

There has been several changes, both to related to cleanness of the maintenance and features.

I did several changes in publishing xsl's to clean up unnecessary functions (migrated with already existing ones), added new functionality regarding to processing of history events (work now better together with Operator), added internal documentation along the principles of XSLTDoc, all referenced documents are now handled through variables. Furthermore,. there were changes to show references on features and routes. Final, and the biggest one, of the changes was a linking between starting points (basic info on article page) and mountain page. This is still somewhat work in progress, as currently only altitude of starting point is brought over to mountain page.

Because references were added to features/routes, css had to be adjusted as well.

Loads of details have been added to Pennine Alps pages.

Several smallish corrections and additions on both Eastern and Western Alps pages as well as in Europe page.

Fixed structural errors in Central Alps section of Eastern Alps page. Added and corrected information on Pennine Alps page.

Well, not so much in the Alps themselves, but the page has undergone some rather substantial changes. Most importantly, Pennine (Valais) Alps is now separated into separate page, undergone plenty of structural changes and had quite a few updates. There have been some corrections and additions on the main Western Alps page as well, but nothing too drastic at this point.

I recently went on to transform Eastern Alps page from old hand-written xhtml format to newer xml source format which is then converted to output xhtml by xsl-transformation. The idea is to get all the pages to same source format which makes it much easier to reuse data, query it and change the appearance should I choose to do so. However, converting the bulk of hand-written (x)html files to proper xml is painstaking to say the least, mostly because my poor choice of not not nesting all elements that would group all things belonging together under common parent-element to save me from writing few extra html-elements per hand.

Thus, xsl-transformation (well, at least not xsl 1.0) is really not up to the task at hand without some serious envelope pushing, possibly involving extension functions. Since this is one time work anyway, since as soon as I have proper xml-documents I can easily use xsl for any format conversions, I started to look at alternatives. Not being a real programmer, I decided on NoteTab Light's clip book feature, which let's you easily script some serious search-replacing using the infinite power of Regular Expression. With all fairness though, I have to say that RegEx sports one seriously steep learning curve. However, once you get the hang of it, it's kicks some serious ass though.

Eastern Alps page is now changed to use pure-xml source file. I also did some reorganizations and added some details.

Plenty of updates on Arctic Norway.

Some changes in css code that apply to event lists.

Created new resource file: tags that is used to tag articles, photo galleries etc in a controlled matter, much of the style of controlled vocabulary. This is the first incarnation of such system to be used on this site, which is more than likely to get revisited sometime in the future. Most likely area to see improvements is handling or hierarchy, now there's none of that.

Since I don't seem to find inspiration to finish what I started with photo galleries, I fixed the gallery page to at least make gallery pages somewhat usable. They haven't been functional for quite some time.

IE7 seemed to have serious issues with span elements containing class attribute within lists. Therefore some changes in publishing scripts to change from dl to ul and adjustments of css files.

Added information on Hohe Tauern mountain page and adjusted xsl-files to output the additional details as well.

Added generation of attribute id in items in list type "item" to facilitate internal linking. Added partial support for hMedia microformat. Small change in content css file.

Since I have recently employed quite a few web 2.0 techniques, I decided to continue the trend. This time around, I added mashup of sorts, namely I exported newly geo-tagged locations from Glockner Group page (Hohe Tauern) to Google Maps map and embedded that map to said Glockner group page.

While I was at it, I also implemented a couple more microformats, this time hCalendar and hAtom. This essentially forced me to employ tagging as well. Next time I feel the need to review something, its likely to mean the inclusion of hReview as well.

Finally, I fixed some bugs in RSS creation (change log), added tiles to index page entries, and changed the presentation of the said entries slightly.

I decided it was time to go more semantic by implementing microformats. In more exact terms, I implemented hCard (html vCard) to item list type, which I use to contain information about huts, hotels, lifts, tramways etc. The big idea behind hCard is to allow browser to recognize contact information on a web site so that it can be easily picked up and exported to vCard, common standard to store contact information that can be exported from and imported into common contact management applications, such as Microsoft Outlook or Google's gmail. Somewhat related, postal address works poorly in the mountains, therefore coordinates are much more usefull in locating huts etc. Especially if you use gps device. Luckily, there's microformat for that as well, namely geo. My own homebrew linking system seems to be not too far away from xfolk So I might change the site a bit so that it takes advantage of that as well.

Forthcoming Firefox 3 and Internet Explorer 8 are likely the first browsers that can handle such microformats out-of-the-box. That being said, there are already plugins for current browsers, eg.

• Tails Export
• Operator

Adjusted map list formatting and added scale and year to be included in the details. Introduced new list type "item" that is used for huts, hotels, lifts, tramways etc. The list is more than likely to get expanded later on to use hCard microformat for more sophisticated handling of contact details, allowing direct merging to vCard compliant address book (such as Microsoft Outlook). Forthcoming Firefox 3 and Internet Explorer 8 are the first browsers that cab handle such micorformats out-of-the-box. That being said, there are already plugins for current browsers.

Further reorganization work and corrections on Hohe Tauern page.

Divided Hohe Tauern page into separate area and mountains pages. Completed and reorganized information.

Several tweaks both in publishing scripts and in css stylesheets to fix issues with the new layout. Most of the pages still use the old layout. Adjusted menu according to recent reorganization of info page. Added rss 2.0 change log.

Like I mentioned earlier, I wasn't too happy with One True Layout layout I had adopted and used on my site for quite a while. Main issues for me were:

• Page clipping when using inline anchors. Obviously very bad, since this site is very big on intra-page anchors.
• Css code used to produce One True Layout is very complex and I don't understand all of it. Take my word for it, not the best position to be in when trying to maintain the site.

Being anal retentive tinker I am, I couldn't go with the solution as it was, of course. Instead I had to roll up my sleeves and incorporate some changes. Most important of them being:

• My version of the layout is elastic all around, rather than using pixel dimensions as was the case in the TJK Article.
• Also, since I think modular css is the best thing since sliced bread, I divided stylesheets into parts:
  • Main stylesheet that contains just layout instructions (that is to say positioning stuff), and not a single instruction to affect formatting (stuff like fonts, colors etc.). There are in fact several layout stylesheets, currently one for properly css-aware browsers and one for less css-savvy7 browsers. Furthermore, there is also separate print stylesheet (@media is extremely handy, btw). Current incarnation of the print layout is not quite what I'd like it to be, mainly because I had to resort to absolute positioning to work. Also, at the moment the design breaks in Internet Explorer.
  • Formatting instructions are placed in separate stylesheet(s) that are imported into layout stylesheets. The beauty of this, is that should I want to change the layout part, I don't have to touch formatting part at all. Much more manageable imho.
• Original version uses well known faux column technique to create equal height columns effect. I didn't want to put mostly transparent background images on right column to finalize that effect in case menu column is the tallest, thus my column separator line between main content area and right sidebar does not scale in case the menu column is the tallest.

To go on and actually implement the new layout, there are some things that I need to do:

• Some images originally intended to be used on black background need to be adjusted in order for them to look nicer on light background
• Have to publish all pages at once or leave old layout styles as well.
• Verify the impact on legacy pages.

Yesterday aside of doing work on info page, I also climbed past few crux issues that stopped me before.

• Bookcat stores descriptions with multiple paragraphs as text strings where line feed character is used to divide string into chapters. The descriptions is also exported like that, so instead of getting the data neatly paragraphed from Bookcat, I get an element with text string containing empty lines. Unfortunately xml parsers don't much care about white space, so that is shows as continuous string without any paragraphs. There would be some solutions that might be a tad easier to implement, but I was persisting on going the whole nine yards, in this that would be reading input (i.e. Bookcat export file) a paragraph at a time and putting each paragraphs into separate child element under main description element. This makes it easy to create proper paragraphing when transforming to xhtml. The trick to do that is to use recursive template that checks if two consecutive line feeds are found, and process conditionally from there. If they are, use substring-before to create first paragraph, and have the template call itself with substring-after as a input string. If no consecutive line feeds were not found, just create a paragraphs and exit. Things are generally easy when you know what to do.
• Same approach worked for getting file part name out of path as well (split by /) which was key ingredient in enabling the same sort of "metadata mining" for processing image files I used earlier for book covers. With some rewrites and few added functionalities, I was able to combine two image processing functions into a single one. The beauty of this is that is now much easier to add new functionality as I only have to change one function. Capitalizing on that, I added support for some more metadata fields. There is still some improvements that need to be done in the final output part (most importantly combining two image processing functions there as well). The last remaining major thing to add in image processing is to add support for images with thumbnail and full res images.

I am about had it with all the issues with One True Layout method. Guess it ain't that true after all. I am likely to switch to some other layout method when I get across to something I like better. Don't expect the layout to change too much though.

Pretty complete rework of info page. Most of the information is the same, but organization is improved and some more information is added. Also some of the dead links have been removed.

Added some details to mountain lists. Few corrections in grading pages. Some improvements in publishing scripts and source documents to simplify the structure and cleanup unused/redundant functions. Some of these changes affect the way how output page header and footer blocks are written to files (gets rid of redundant divs). Added few more instructions in css files which made it possible to remove some markup from source files. Plenty of changes in how source files are stored.

ABC goes Ajax. I added nice javascript code tablesort.js which uses Ajax technology to turn static tables into sortable ones. While I was at it, I also tweaked table css a bit to produce nicer looking table. Also, I couldn't resist adding title attribute to grade link to show link description as tooltip. All this is currently in action on 4000m peaks in the Alps.

I seem to have been in very productive mood regarding to coding of the website during last few days. Today I solved the crux in regards of using Bookcat to manage book information, namely inserting the images. The solution involves:

1. Import book images stored by Bookcat into image cataloguing software, that is able to export image dimensions (exif data) into xml file together with file name.
2. Use xslt file to match cover image filename (which is part of the book details exported by Bookcat) against image xml-file using file name as a key.
3. Copy over the image details as needed.
4. Since Bookcat does not limit the cover sizes in any way (and those vary a great deal), I wrote few lines of extra code in xsl to "scale" the image dimensions. Well, it does not really scale a thing, it just compares image dimensions against defined max. width and height, then sets calculates new values for width and height attributes so that the image does not exceed the set max values and original image aspect ratio is retained.

Fixed few bugs and typos in css scripts. Added rss logos on index page links.

Related to changes done earlier on index page, a number of changes in publishing scripts to fixes, improve and add functionality of the publishing scripts. Also some minor tweaks in css stylesheets. Also updates some of the links on index page right column.

Yet another major change on the technical side. Up until now index-page has been manually maintained without it taking any advantages of auto-magic that takes place at the publish time courtesy of xsl transformations. That's all in the past now.

Further lots of changes in source xml files. Minor tweak in css stylesheets. Do-over of Mont Blanc Massif page with lots of added information and reorganization.

During last few days there has been lots of changes in source xml files. Many of these changes are not directly affecting he output but due to more uniform structure of source files I have been ably to simplify publishing xsl-files. While doing that I also added some missing functionalities to xsl files. I also tweaked one of css files a bit so that output html looks slightly better in some scenarios.

Added few routes to both alpine and ice grading pages. Additions to Scandinavia.

Added lots of details to Lofoten and Tysfjord (Stetind). Added information to Himalaya page. Adedd new routes to alpine and ice grading pages. Corrected typos on Caucasus page. Some changes and additions in publishing scripts to make them compatible with few shortcuts in source files. Also added some features to outputting book data, now books with partial details are handled better with little or no unnecessary punctuation. Changed css for print media so that it now uses almost all definitions from screen media stylesheet, exception are definitions for menu-column, links, content-column width and background (currentle there seems to be problem in handling link width in Firefox 2.x, works fine with IE7).

Added details to alpine grading and Himalaya pages. Plenty of changes in Mont Blanc massif page (grouping of peaks).

Small change in print media css that makes the site print much better with IE7. IE7 seems to interpret width 100% to be related to the actual page width, not considering margins, thus it makes printing entire contents impossible. Width set to * saves the day.

Plenty of added information and restructuring in Himalaya page.

Lots of changes in css stylesheets for easier management (none should affect the actual layout). Few changes in index-page to make it use same style names as the rest of the site to get rid of unnecessary duplicate styles.

Some changes in publish scripts (normalize_string -template) and css-stylesheets. Added static news to the front page, this might get replaced at some point by ajax news.

Lots of changes on Romsdal page, now mountains have been separated from article part. Also now Romsdal mountain page is pure xml. To facilitate this, there are numerous changes in publishing scripts. Also, the page has feature data level under mountains, there are several changes in css-stylesheets (mainly of contextual kind) to make pages with features more readable. New model is not quite final when it comes to functionality, but should be about as usable as the old model at the moment. There are also some changes in Caucasus page content.

Lots of changes in publishing stylesheets to use more attributes instead of elements to produce more compact and readable files. Also redundant removed redundant elements. Added new features such as map list and "pass-through" of some xml structures from source to output files (normally those would be processed against database-files. This is done so that is easier to test some other changes and also to allow gradual migration of current static files into data-driven pages. Also minor corrections and tweaks here and there and more defaults and conditional processing applies. It is surprising how easy it is top spend few productive hours of work with nothing visible changed at all!

Changed publish script so that it now can process more nested chapters then before (previously three levels, now five). Current solution is not very elegant though, so it possibly be changed later on. Changed navigation system for mountain pages slightly. Conversion of mountain pages to proper xml has been started (all pages converted partially). This is a big job that has to be completed before there's much sense in starting to implement further development on the publishing side.

Adjusted publishing xslt so that it can n ow properly sort dates either in ascending or descending order using substring-method (source format dd.mm.yyyy).

Made few corrections in cgi-scripts and moved one script into subfolder for it to be able to use same html header and footer files that are used by other scripts. Customized search script so that it now load new search form from external file.

Added xmlns="http://www.w3.org/1999/xhtml" in stylesheet used to output html pages to output xmlns declaration in html root element.

Added details to several areas under Nordic countries. Separated Romsdal from main Nordic countries page. Some changes in publishing stylesheets.

Added details to Hemsedal and added two new areas, Laerdal and Valdres in Nordic countries.

Changed several mountain pages so that they are now using xsl page to include common html formatting. This is still far from complete solution as the pages now include well-formed (but not necessarily valid) xhtml data, that is included as is between html header and footers. The principal limitation of this is that the xhtml data is still completely static with no data driven parts. However, even this limited solution is significant step-up compared to former situation where most mountain pages were completely staticc and did not make any use of common formatting scripts. Plenty of information added to Siteinfo.

Changed several mountain pages so that they are now using xsl page to include common html formatting. This is still far from complete solution as the pages now include well-formed xhtml data, that is included as is between html header and footers. Even this solution makes maintenance of the page far easier and once the transfer of all mountain pages is complete, it is finally possible to start changing the pages into real xml-pages gradually. Most importantly, as all the transferred pages are well-formed, it should be much easier to write xsl-stylesheets to transfer those pages into proper xml automatically or automatically, manual transformation would be exhaustive as some of the pages are huge (way more than 100kB, that is quite a bit considering they are text-only). Corrected link on info.htm so that it refers to correct page.

Two pages converted from html to xml. Also minor additions and correction to stylesheets used to convert xml into xhtml. Corrected paths referring to graphics on gallery and siteinfo pages.

Changes navigation script (dTree 2.05) so that currently a single script can be used independent of relative path to images used in menu as path is given as parameter on a page calling for navigation tree script. Moved few photo pages from under the mountain under gallery.

Since Internet Explorer 6 doesn't support alpha channel transparency in PNG-24, replaced transparent PNG-24 images with PNG-8 images (site logo graphics).

Quite a lot of changes mostly due to reorganization of graphics, stylesheets etc. Currently all the resources related to layout (some of the graphics, JavaScripts and CSS stylesheets are combined under layout root directory so that there one grouping folder (theme). This way management should be easier and it should be possible to adjust layout easier than before. Also, some of the gif graphics were replaced with similar PNG graphics, some graphics were moved to different location and orphaned graphics were deleted. Also, some typos were corrected and "meta http-equiv" was added to some pages that still missed it.

Rename himalaya.htm to asia_himalaya.htm, westernalps.htm to alps_west.htm, easternalps.htm to alps_east.htm. Separated Caucasus from asia.htm, currently the page is asia_caucasus.htm. Changes links on all pages to refer to new pages. Several corrections on Caucasus page (spelling, markup and information. Adedd and reorganised info on Caucasus page. Added information about guidebooks coverings European Alps (europe.htm).

Added details to Caucasus.

Added details to Caucasus.

Further improvements in xhtml 1.1 compatibility (break out of frame currently uses simple Javascript instead of target="_top" and several adjustments in forms used on siteinfo page). Currently all xslt-generated pages are xhmtl 1.1. strict compatible apart from alpine grading page (that uses several non standard attributes in span-elements. Those elements do nothing but are supposed to make eventual transformation into data driven grading pages easier). Next step of the site development will probably be go-through and modularization of xslt-files used to transform data into pages and css files to get rid of redundant code and to make them easier to understand and manage. Also, I am not happy with "One True Layout" css layout schemes bug that leads to highly annoying behavior when using anchors to jump into predefined spots within an document.

Changed blockquote to p class="quote" for xhtml compatibility. Several corrections on many pages and few corrections to xslt files for better xhtml compatibility (majority of pages created through xslt are now fully xhtml 1.1 compatible; however mountain pages are still hand coded and mostly not valid xhtml 1.1 files). Corrected xslt used to create recommended page.

Adjusted index.xhtml.css to get rid of unwanted margins and centered the page

Finally found the reason for persistent character display problem with extended characters. It seems that (x)html-documents have to have "<meta http-equiv="content-type" content="text/html; charset=utf-8"/>", otherwise Browser (IE) does not switch to UTF-8 even if the file has "<?xml version="1.0" encoding="UTF-8"?>". Changed photo gallery page to use same layout template as rest of the pages.

Added several routes to alpine grading page. Added new span elements in Alpine grading page (NCCS). Those have currently no impact whatsoever but will facilitate new functionality later. Some changes in Jotunheimen page.

Added new grouping element under mountain. This is used to group routes according to their geographical location (mostly faces). Also used to store common info such as access, general roiute character etc. This structure currently only applied in Romsdal (Nordic countries). Changes in CSS files to present the new grouping. Also changes in html grouping elements (div) that could facilitate additional functionality later on. Added several Bolivian climbs in Alpine grading page.

Added info for Cascades (North America).

Added info for Himalaya page (mainly Khumbu area).

Added routes to alpine grading.

Added info for Romsdal (Norway). Especially several winter routes.

Added info for Romsdal (Norway). Especially several winter routes.

Corrections and additions to Cordillera Blanca (South America).

Corrections and additions to alpine grading.

Corrections to alpine grading.

Added few routes. Corrected misgrading on Fivarute (Store Trolltind).

Added details to Nordic countries. Added few routes to grading table. Added Norwegian over4all grade to alpine grading.

Completed info for New Zealand alpine grading (new routes, details added).

Added info for Europe and Hohe Tauern. Reorganized Nepal. Added several routes and details into alpine grading page.

Reorganization of Hohe Tauern. Completed info for Hohe Tauern. Few additions to alpine grading.

Reorganization of Northern Andes, Puna de Atacama and High Andes. Completions to Cordillera Blanca and Cordillera Huayhuash (all South America). Adjusted style sheets.

Reorganised Southern Peru & Bolivia. Added info for Cordillera Blanca, Cordillera Huayhuash, Bolivia and Patagonia (South America).

Added info for Cordillera Blanca and Cordillera Huayhuash (South America).

Added info on Europe (Tatra, Pyrenees). Reorganised southern America (Cordillera Blanca, Cordillera Huayhuash).

Reorganised North America page. Added info on route information organization (Mountains home page).

Added info for Venediger group in Hohe Tauern. Added and corrected links in Grossglockner page. Adedd much information into Jotunheimen page. Jotunheimen mountains currently not linked at all.

Added details to Glocknergroup. Reorganised and added info for Hohe Tauern. Added info for Ankogel group. Adedd routes to Hochalmspitze. Separated Hohe Tauern from Eastern Alps. Cleaned info on Grossglockner page (still under construction).

Updated new mountains, routes and info in Glockner group section (Eastern Alps).

Filled comment into anchor (a) elements. This circumvents empty anchor element inconvenience in Firefox 1.5x. Changed Equipment guide (Finnish) to use same navigation tree than the rest of the site and simplifield Equipment guide frame structure. Deleted unneeded javascripts, images and pages.

Applied xhtml layout to Mountain pages (Done manually, done this time to make it easier to transfer mountain pages gradually to publishing system used for other pages). Added rule in print css which should print link urls in parenthesis after link title in brosers that support pseudoclass after. Set side margins in print stylesheet. Corrected links in Alps map that appears on Europe page.

Applied a hack (&nbsp; with line height of image height) to make IE show background image on empty div on the index page. Modularization of stylesheets (style sheet for print separated into separate style sheet to keep style sheet easier to understand). Also layout portion is separated page. Applied One True Layout based layout to all xhtml pages. This should solve column height problems. Printing extra pages because of this should also be eliminated. Changed publish scripts to output two separate links to xhtml, one for media="print" and one for media="screen". This circumvents Firefox float inconvenience (bug?) when printing. Downside is the need to include all the elements in print stylesheet separately, if @import is used floated elements mess printing in Firefox. Corrected few werrors in xslt-files. Added new feature in publish scripts that allows embedding chapters from external files and adds link to rest of the file. This is use dto link grade document from Info. This will also be used for mountain pages.

Redesign of index-page. Now uses css div based layout instead of html tables. Changed event list sorting to sort by date (uses substring function to split date string into year, month and day).

Changed default_xhtml.css. Changed Siteinfo organization a bit and added link to Change_log. Added new list type (event list) in source documents.