Unfortunately, This (good) tutorial is outdated on many points.
It has been wikified and updated (a year and half ago...) by Tom Roche here.

In some (simple) cases, you should consider using the "translator framework".

I use Firebug to get XPath.

I'd actually put this differently. In 99% of the cases addressed by Adam Crymbel's tutorial you will be _much_ better of using the translator framework.

So below is my first attempt with the translator framework. It seems very nice! Some questions etc:

- The field Scaffold/Target is not used by FW. Is it used at all?

- What should be put in Scaffold/Min. Version?

- How can I normalize the date (ISO format preferred)?

- Why does not my "tags" work? Could Scaffold perhaps tell that the object type is wrong?

- What is "libraryCatalog" and where did I specify that? ;-)

- Any suggestion for what to do with FW.MultiScraper and how to block list pages for FW.Scraper?

Tested on this page:
http://www.psychologytoday.com/blog/mad-in-america/201112/cognitive-therapy-found-effective-in-unmedicated-psychotic-patients-and-o

Here is the FW code:

function detectWeb(doc, url) { return FW.detectWeb(doc, url); }
function doWeb(doc, url) { return FW.doWeb(doc, url); }

FW.Scraper({
itemType : 'blogPost',
// detect : FW.Url().match(/(\/node\/\d+|articles\/[a-zA-Z0-9]+\.shtml)/),
detect : FW.Url().match("^http://(?:www.)?psychologytoday.com/blog/"),
title : FW.Xpath('/html/body/div/div[4]/div/div[2]/div/div/h1').text().trim(),
creators : FW.Xpath('/html/body/div/div[4]/div/div/div/div[2]/div[2]')
.text().replace(/\s/," ")
.trim()
.replace(/^by /, "")
.cleanAuthor("author"),
date : FW.Xpath('/html/body/div/div[4]/div/div[2]/div/div[3]/span')
.text()
.match(/[a-z]+ [0-9]+, 20[0-9][0-9]/i),
// tags : FW.Xpath('//div[@class="terms terms-inline"]/ul/li/a[@rel="tag"]').text(),
tags : FW.Xpath('/html/body/div/div[4]/div/div[2]/div/div[4]/div[3]/div[2]/div[4]')
.text()
.replace(/Tags:/, "")
.trim()
.split(/,\s*/),
blogTitle : FW.Xpath('/html/body/div/div[4]/div/div/div/div[2]/h1/a').text().trim(),
abstractNote : FW.Xpath('/html/body/div/div[4]/div/div[2]/div/div[2]').text().trim(),
downloadAssociatedFiles : false
});

Please do direct further questions to zotero-dev, but I'll answer the Scaffold-related ones here for good measure. If you do not insert the framework code manually into the translator by pasting it from Erik's repository, the framework code will be included if and only if the checkbox is checked. There is a little bit of boilerplate-- the detectWeb and doWeb functions -- that will be inserted and not removed when checking and unckecking the checkbox, but the real framework code is a massive minified line that is blissfully hidden from you while you are working within Scaffold.

Scaffold will handle the date automatically, and don't worry about the minimum version; if it matters, I'll set it before putting it in the repository (it almost never matters).

Thanks adamsmith, I will go to zotero-dev. And thanks for the hint about abbreviating the paths.

Thanks ajlyon, I see. Perhaps some message about what is included/removed when checking/unchecking that checkbox would make it more clear?

Info about testing in standalone version should I ask that here, or?