First things first: if you don’t know what I mean by “TextToHTML “, you won’t get much from this post. In this case, I recommend that you start with this introduction.
In short, the TextToHTML script has two roles:
1/ find HTML strings in a SharePoint page
2/ Convert these strings into actual HTML

Yesterday, I came across a discussion between @EUSP, @webdes03 and @ebrackley on Twitter:
“TextToHTML can be a bad performer when tasked with lots of HTML; also research DVWPs”

This is a timely comment, as I am about to present at SharePoint Saturday EMEA. Let me expand on this, and provide a few hints on how to better use the HTML Calculated Column.

Remember: TextToHTML is not the only way

Right, most of the examples you’ll find on my blog rely on the TextToHTML script, embedded in a Content Editor Web Part. But there are other ways to render the HTML. In particular, the Data View Web Part, used in crosslist mode, or the Content Query Web Part can directly do the rendering. For more information, see these series:
– for MOSS
KPI roll-up in MOSS
– for SharePoint 2007 (applies to both wss and MOSS):
KPI roll-up in SharePoint (Part I)
KPI roll-up in SharePoint (Part II)

The latter will be the theme of my presentation at SharePoint Saturday.

Get a faster browser

Not always a choice, but if you can, upgrade your browser to the latest version. These days, browsers are improving by leaps and bounds, with Google leading the charge.
My tests show that the TextToHTML script is four times faster in IE 8 than in IE 7. That’s huge! Some other browsers offer even better response time.

Faster TextToHTML

If you are using the TextToHTML script, you’ve certainly got the current version from the download section. What I am making available over there is a generic script, for both SharePoint 2003 and 2007. If you are on SharePoint 2007, you can make it more specific and thus faster.

Focus on the main content

As is, the script will scan the whole page to find table cells (TD elements):

TextToHTML(document.getElementsByTagName("TD"),regexpTD);

This is a waste of time, as you just need the main content, excluding header and Quicklaunch. You can easily do this by replacing the above line with the following code:

var theMainContent = document.getElementById("MSO_MainContent");
TextToHTML(theMainContent.getElementsByTagName("TD"),regexpTD);

Use selectors

Instead of grabbing all the cells in the document, try to restrict your scope to the cells that may contain HTML Calculated Columns.

For example, Paul Grenier proposed a jQuery version of my script, which allows you to grab only certain cells. If HTML Calculated Columns are only in List View Web Parts on your page, the simple $(“td.ms-vb2”) selector should be enough. For better performance, you may want to combine it with my first advice, and focus on the main content: $(“#MSO_MainContent td.ms-vb2”).

Here again, a modern browser will give you better performance (for example if it has native tools to find elements by class name).

Place your script close to the point of consumption

Imagine that you are displaying 10 lists on your page, but only the third one is using HTML calculated columns. If you place the CEWP that contains TextToHTML at the bottom of the page, the script will go through the 10 lists. If you place the script right below the third list: when the script runs, it will be faster as it only sees 3 lists, the 7 others being displayed later.
btw placing the scripts close to the point of consumption is a good practice that also applies to other scripts, not just here. In traditional Web design, you would find all the scripts under the head section of the page, but this practice has evolved now that pages are more dynamic.

If you have other ideas, feel free to share them with me and the other readers! I’ll push some of these performance improvements in the next release of the TextToHTML script. I am also looking for volunteers to test the beta versions…