Column: Stop Reading the City’s Website
Among the incidental, minor topics touched on at the April 16, 2014 mayoral candidate forum was the city of Ann Arbor’s website. Praise was not heaped upon it – as a2gov.org was described by one candidate (Sabra Briere) as “a terrible website to try to tell anybody how to navigate.” That’s not an uncommon view.
How to search just the city of Ann Arbor’s website with Google and reduce frustration when you are unable to find information by navigating to it.
So stop navigating it. Stop “reading” it.
Start searching it – and you’ll probably find what you’re looking for.
That’s not to defend the user interface or the look and feel of a2gov.org. Maybe it is terrible. I don’t have a strong opinion on that. It’s worth noting that the city council approved a contract with Keystone Media ($26,900) at its Oct. 21, 2013 meeting to redesign the basic templates for the city’s website.
I’m not sure if that work has yet been implemented – as I just don’t pay that much attention to the look and feel or the navigational features of the city’s website. That’s despite the fact that part of The Job is to look stuff up – quite frequently on the city’s website. And mostly I find what I’m looking for pretty quickly.
So my point in writing is to share one simple technique I use dozens of times a day to do The Job. I use Google search – but constrain the search to just the one website where I’m looking.
The syntax goes like this:
site:terrible-to-navigate-website.com word-1 word-2 word-3 word-4
To illustrate how effective this can be, let’s consider an example from real life. Last year, the March 18, 2013 city council meeting featured extensive public commentary on the 413 E. Huron project that pushed the meeting conclusion to around 2 a.m. I wanted to include in our meeting report the names of all the speakers.
But the final speaker stated her name in an unclear way, so that it was not possible to discern what it was – even after listening to an audio recording several times. She did, however, identify herself as a member of the Old Fourth Ward Historic District Study Committee.
That historic district study took place 30 years ago. What are the chances the city’s website had a list of the committee member names to which I could compare the audio? To find out, I didn’t read the city’s website. I did not navigate the city’s website. I did not even “visit” the city’s website.
The first thing I tried was using Google’s site-specific search. Here’s a screen shot of the search results from Google:
Screenshot of search results from “site:a2gov.org historic district study committee for the Old Fourth Ward”
The third result looked likely to have a listing of committee member names. It did. Comparing the audio to the list was easy work: Cappy Bilakos.
I don’t mean to suggest that site-specific search syntax is magical or that the city should not worry about improving the readability and navigation for its website.
But I would suggest that frustration levels for users could be reduced if Google’s site-specific search were used as a primary strategy for finding information. That goes not just for the city’s website, but for the Internet in general.
The Chronicle could not survive without regular voluntary subscriptions to support our coverage of public bodies like the Ann Arbor city council. We sit on the hard bench so that you don’t have to. Click this link for details: Subscribe to The Chronicle. And if you’re already supporting us, please encourage your friends, neighbors and colleagues to help support The Chronicle, too!
I’ve thought that the Washtenaw County website has always done a great job of making information accessible. The City website is quite disappointing in contrast.
One of my pet peeves is that all city announcements are in pdf form. For short-term announcements (like “waste and recycling pickup will be moved ahead one day this week”), there should be an announcements box in simple text that doesn’t require downloading a pdf in order to read it.
That’s exactly what I do too.
But speaking of site search, the Chronicle’s search box stopped working for me several months ago. All I get is “Loading” and no results.
Re: [2] Jim, sorry about that. Can you send along the browser and hardware combo you use and I’ll ask our web guy to have a look to see what might be going wrong. (It works for me. It’s just Google’s site search – with some filtering on duplicative results – piped into a Chronicle template.)
Like Mr. Rees, I would get the unending “Loading” message when using the Chronicle’s search feature. I use the FireFox browser on a MS Windows 7 platform, with a variety of security add-ons to protect privacy while browsing.
One of my security add-ons, Ghostery, prevents data gathering by web sites. I found that configuring Ghostery to allow the “Google AJAX Search API” to collect private data, the Chronicle’s search function works as it should. On the other hand, if I go to google.com and type in “annarborchronicle.com, search.term” there is no need to disable any protections.
Similarly, I have found that I must enable something called “Echo” in order for the mLive.com comments to work in my browser.
On the web, we are just the aggregate data that has been collected about us.
The City of Ann Arbor doesn’t publish everything on its web site. Rather, it uses sites like Legistar for its meeting minutes and agendas, and Municode to publish its city code. So even a complex search like Dave describes won’t turn up everything.
If you do search Legistar (http://a2gov.legistar.com) you will find that you can locate the text of meeting minutes and agendas, but the attachments and files associated with those agenda items are unsearchable.
I’d put the blame squarely on the City of Ann Arbor CIO, but that position has been vacant for a while. Anyone know what’s up with filling it?
Legistar search tips, which I wrote up for Damn Arbor last year: [link]
I need to correct my earlier complaint – I just went to the website and found that they are now displaying some short announcements as text. Good move!
Re: “… but the attachments and files associated with those agenda items are unsearchable.”
While Legistar does not return attached files separate items in the search results, I _believe_ Legistar searches of Legislation are sensitive to the content of attachments. That is, if an attachment has a target term, it will return the agenda item to which that attachment is attached. If you search Legislation for February 28, 2014 it will turn up this item: [link]. But that’s because that string is in one of the attachments, not because the string is in the body of the agenda item itself. Still you’re give no clue which attachment. So you have to open up each attachment and check. So Legistar search should be improved so that the objects that Legistar returns in its search results include attachments, not just agenda items.
8) That would be nice if it were true, Dave, but it isn’t.
Do a search like this, for one of the members of the Downtown Area Citizens Advisory Council, whose name is in the minutes of the meetings on this week’s agenda:
site:a2gov.legistar.com “Stephan Trendov”
You’ll get zero results, even though his name is in the attachments to 14-0572
[link]
Correction in (9), Trendov was a visitor listed at that meeting, not a member of the DACAC. The note on the unsearchability stands.
Ed, you might have assumed that Legistar treats quoted strings as exact matches, but I don’t think it does. I think it includes the quotes in its search. If you just type in Trendov, you do get 14-0572 as one of the results: [screenshot] So, as I wrote above, using a different example to illustrate, it appears that Legistar search is sensitive to the contents of attachments to agenda items.
Re: “Ed, you might have assumed that Legistar treats quoted strings as exact matches, but I don’t think it does. I think it includes the quotes in its search.”
Well, that’s obviously wrong, because Legistar returns 14-0572 as a result even with the quotes.
What I missed is that you were doing a Google site-specific search on Legistar, and your complaint is not about Legistar’s search, but about the invisibility of attachments to Google. So perhaps Google doesn’t want to look any deeper than the agenda items, or is not able to look deeper.
I meant to be describing Legistar’s own search box – which is what I use when I’m looking for something on Legistar. It lets you sort results by “final action” which is sometimes handy. And you can narrow the scope of the search by year at the outset, if you have an idea of what year you’re looking for.
You got it, Dave.
Council produces a set of PDFs and other documents in support of a meeting. Those documents are hidden in Legistar, out of the eyes of Google search. You can’t even download all of them in any straightforward way that I have seen and host them yourself.
I take that back – there is some code on github called “Councilmatic” that eats everything in a municipal Legistar system and provides a modern user interface for it. See [link]
or
[link]
It’s running in Philadelphia.
Re: “Those documents are hidden in Legistar, out of the eyes of Google search.”
I don’t think this is quite as egregious a problem as you’re describing it. As I’ve shown, attachments are in fact visible to Legistar’s own built-in search. They’re only “hidden” if you decline to use Legistar’s built-in search. Is it perfect? No.
For the use case of an Ann Arbor citizen who’s interested in following the workings of local government, I think needs are reasonably well, but not perfectly met. (E.g., it would be better if Legistar returned attachments as objects in the search results.) Yes, for the use case of a researcher who’s investigating legislation across several communities, then Google is going to miss Ann Arbor’s Legistar attachments. Perhaps there are other, more common use cases whose needs are also not well served.
As I see it, either Google does not want to index any deeper than the agenda items themselves, or Google is not able to do so, because of something about the architecture of Legistar’s information. Maybe Google should be blamed, or maybe the software engineers who made Legistar should be blamed, or maybe there’s actually some simple straightforward switch that can be thrown in the city’s implementation of Legistar that will invite Google’s spiders to index the attachments.
There are certainly other aspects of Legistar that could be improved. An example that comes easily to mind – because we’d like to have icalendar data feeds for The Chronicle’s event display – is the fact that the Legistar calendaring module does not generate a URL for icalendar compliant event data. As a result, someone at the city has to do double data entry for Legistar events and for the Google-based event calendar maintained by the city.
My understanding is that the city is currently working with Granicus, the entity that now owns Legistar, on some upgrades and improvements to the city’s implementation – while mulling the possibility of eventually changing vendors. The potential step of changing vendors, as I understand it, is probably a very long way off – and that would bring along with it the challenges of migrating all that Legistar data into a different piece of software. But if you think that visibility to search engines of all information stored in the system should be a priority, then the time is probably ripe for conveying that to the city.
Dave, I just think it’s curious (and unfortunate) that to get the most out of city web sites, that you have to know so much about each of their failings to know that you are supposed to search one of them with Google search because the on-site search is so badly broken, but for the other one you’re supposed to search it with on-site search because the Google search is so badly broken. Hardly inspires confidence!