And one of the endless reasons I respect Jon is his embodiment of being the antithesis to the Silicon Valley holy devotion to the cutting edge, is his devotion to the other edge:
I’ve spent my whole career exploring and explaining many of the technologies that enable — or could enable — networked education. And while I was often seen as an innovator, the truth is that much of my work happened on the trailing edge, not the leading edge.
There’s a reason I keep finding novel uses for these trailing-edge technologies. I see them not as closed products and services, but rather as toolkits that invite their users to adapt and extend them.MOOCs need to be user innovation toolkits (Oct 11, 2013)
I have one in my arsenal I use so regularly I do not notice it anymore. And I was reminded of this in, again, the most unlikely ways. It came from a tweet that I spotted and raised my curiosity about the idea of a del.icio.us comeback (one of the early web technologies I so miss now.
I thought Marshall’s followup tweet was related to this, and loathe as I am to go in LinkedIn (long story), I followed the link
Marshall’s blog post LinkedOut from LinkedIn references a technology I not only used long ago but I totally forget is built into my own blog. It’s Google’s Custom Search Engine (CSE). It sounds very edgy indeed, but think of it as using Google Search muscles to do it’s work not on everything it indexes, but a selected list of sites- ones that you choose.
Why not search all that the vast Google Eye scans? Well for many things, it’s more relevant and useful to search from a more curated set of sources.
I am pretty sure I learned of this tool from Scott Leslie and remember first putting it to use when I was working with NMC and found it a way to provide a search selected resource sites to feed ideas for the Horizon Project. I hacked some code to integrate it into the wikis used back then.
The NMC sites have all been mothballed, but you can find shreds of the Horizon Project Wiki’s in the Wayback machine there linking to the embedded Google Custom search. You get two old technologies for one:
We are harnessing the power of Google to provide you a keyword search from a selected set of major technology web sites. As you discover resources relevant to this horizon project, do not forget to tag them in delicious.com!
At some point on this blog I decided to make use of Google CSE just to search cogdogblog.com – maybe it was a recognition that the built in WordPress search often did not let you zero in the way you can using the various Google Search operators.
When was it? Hmmm, a quick search with the tool I am talking about, and yes, it was March 10, 2010 that I scoped out how I dug into the search form of CSE I set up to search just cogdogblog.com and embedded it into my blog. And added bonus, as I noted, is that it not only searches this blog, but anything else on the cogdogblog.com domain, like a pile of code or some of the web-based presentations or other junk I summarize as “stuff“.
I have it somewhat duct-taped into cogdogblog.com via a custom php file that sites at the top of my domain. But my clever clever hack is that I have the whole CSE set to go from a browser bookmarklet tool (talk about trailing edge!) I made long long ago a web tool where you can make one of these that can issue a search from ANY WordPress site you select (just knowing a search for “blah” is issues by taking on to any wordpress blog
?s=blah Try it! http://cogdogblog.com/?s=blah
But here is the thing- the browser bookmarklet tool lets me do that from anywhere I am roaming in the web. I do not have to be at my blog. I can either invoke it and type the keywords in a box.
If I am say, reading a post on Jon Udell’s site, clicking the tool lets me search my blog from there:
But even better, if I see a phrase in Jon;s site, I can just highlight it, and when I hit the bookmark, it runs a search on that phrase in my blog.
Again, if you use my WordPress Bookmarklet Maker Tool (created way back in 2005), it will provide this functionality by passing the search terms to the built in WordPress search. This is insanely useful (well I think so).
My insight after setting up the Google Custom Search engine for cogdogblog.com, was to modify the bookmarklet code to send the entered or selected search terms to my CSE. And this has been the way I regularly search my own blog, I rarely use the WordPress search.
I had really forgotten even that this is the way it works! I decided this week to tinker and improve it, especially after realizing that among all the things Google has sent to its graveyard, CSE still works (it’s been around since 2006!).
The first thing was making a front end access to this form (I mostly use the bookmarklet tool) so I could make this into a demo. A new trick I found is a way to have it display popular queries that work as links:
I found this in the stats section of my CSE (not linkable). I am not sure who entered many of these.
So if you want to see my posts about
cse, just enter it in the form and voila, you can see the results I have been seeing since 2011 – it’s rendered through a funky PHP script I figured a way to set up that uses WordPress features for calling headers, and templates, without being part of WordPress (this is very specific to my theme, but this exists at a custom php at the root of my WordPress site:
/* Set up WordPress stuff */
<?php get_template_part( 'template-parts/wrapper', 'top' ); ?>
<div id="primary" class="content-area">
<main id="main" class="site-main" role="main">
<?php get_template_part( 'template-parts/cover', 'postbones' ); ?>
<?php get_template_part( 'template-parts/content', 'foundbones' ); ?>
</main><!-- #main -->
</div><!-- #primary -->
<?php wp_footer(); ?>
It tells basically creates a setting that says “don’t use the WordPress template flow, but use everything else”
I have the results set up so it offers via a link, a way to see the same search through the standard WordPress search, and it sets up the form so one could modify the search and go again.
Now below this are ads. I found no way to suppress the ads, so I am feeding the machine. But after that are all the results.
Now in my review of CSE this week I found something new- I could set a flag that would let me see the results applied to images on my site. There is now that toggle tab where you can flip from the web site results of the search, to see images that match.
Now I am not too sure how google is making the inference (I might guess it’s images from blog posts that match “cse”?) but woah, I now have a way to use google search to find images in my blog
I did not expect to find this!
Now one more trick to make the circle complete. If you follow the link for the standard WordPress search results
I added a link in the header under the number of results, that can send me back to the CSE search. So I can get there in multiple ways now.
This may seem like a whole lot of smoke and mirror mumbo jumbo. but I can tell you I use all of this sometimes several times a day. To recap, I can use:
- A browser bookmarklet tool (this goes back to 2001) created with a tool I made in 2005 so I can search my blog from any where I am online, even triggered from selected text
- The results are using a Google customized search tool (circa 2006) to find things just in my domain.
- If all goes well, I might be able soon to bookmark results to del.icio.us!
And all of this works like a sharp axe in 2021. So go ahead and gloat about your AI infused semantic blockchain… I’ll keep applying the trailing edge tools, and will sharpen them as I go.