Caspio’s Lessons
June 29th, 2008 | Published in Car Tools, Journalism | 6 Comments
Been awhile since I wrote about Caspio, and since then they’ve only gained more media clients, which I suppose could be a lesson for me. But I think not. Rather, I hope what we’ll see in the months and years to come are the lessons that Matt Wynn offers from his experiences using Caspio. Here’s your nutgraf: “My conclusion on Caspio is that they do one thing very well. But other, cheaper alternatives do it just as well. Further, to learn to make it do otherwise seems pointless, especially seeing as we would be paying for the luxury of learning to hack it.” (The emphasis is mine.)
Caspio’s David Milliron spoke at this year’s Special Libraries Association conference at a panel organized by SLA’s News Division, which includes many newspaper and broadcast librarians. It’s easy to see why: a lot of these folks are being asked to do new things, to be more involved with their organization’s Web sites, and to do it with fewer people. Seems like a pretty good opportunity for Caspio, and I don’t fault them for recognizing that. The problem I have is that the promise of Caspio is in the short-term; no matter how many features they add (my personal favorite being the Data Sheet Find and Replace one: “You no longer need to export your table outside of Caspio Bridge for this type data modification.”), you’ll never get the flexibility and control over your apps that you do when you build your own stuff. Despite what Milliron says, there are very real and serious differences between Caspio and Web application frameworks.
Maybe that’s the real lesson that journalism folks need to heed: that the costs of learning Caspio go beyond the monthly fees and the potential cost of switching to another tool (having to re-do your existing apps). Caspio is, as Matt says, good at doing some pretty basic stuff when it comes to putting data online. But if you want to go beyond Ye Olde Data Ghettoe, you’ll have to learn some programming anyway. So why learn something that can only be used on a closed system that you have to rent? Matt’s alternative happens to be PHP/MySQL based, but he’s not going to be paying for using either of those. And if suddenly MySQL decides to charge corporate users or something equally far-fetched, he can switch to Postgres or SQLite without starting from scratch.
I realize many, many folks in newsrooms can say, “Um, pardon me, but we don’t have a Matt Wynn.” Or maybe you do, but he’s insanely busy all the time. That’s a very common situation. But the real long-term question is this: if your organization is never going to want to do anything more than put up isolated search pages serving up content that no search engine can reliably find, you’re still gonna pay every month for that privilege by using Caspio. And if you hope and plan on doing more someday, even if that’s not today, then you’ll have almost nothing to transfer to that effort by using Caspio, since one of their chief claims is that you don’t have to learn any programming to use it.
So if learning more is a part of your plan, why not spend the time learning a system that doesn’t charge you for that time? By adding Caspio experience to your resume, what real skills have you gained aside from the ability to point and click?
June 29th, 2008 at 7:01 pm (#)
Dismissive, vaguely angry, off-point comment from Milliron in 5 … 4 … 3 …
June 30th, 2008 at 10:52 am (#)
“…adding Caspio experience to your resume…” Frightening that this has become an acceptable — if not desired — skillset, isn’t it?
At $8/mo./datapage, there are other options available. I think organizations aren’t even considering what they want out of Caspio when they sign up. And that’s causing them to overlook other better, cheaper solutions.
Having ten datasets up will cost around $1000 every year. If all a news org needs is a search box/results container, that money could be funneled toward an Elancer who could build a tool to do just that. If you want to throw in maps, add it to the job description — it’s definitely doable.
If you want to try pushing Caspio to the limits, then you have to be willing to hack around, anyway. Use the cash investment to send a curious staffer to a PHP class at a community college, or shoot them off to the Django classes at IRE/NICAR.
I guess what bothers me most about Caspio is that its business plan is preying on a short-sighted systemic mistake endemic at newspapers. It’s ludicrous that money — which could go towards actual journalism — has to be diverted to build a workaround over a purely internal problem.
June 30th, 2008 at 1:55 pm (#)
Oooh wee, it might get ugly in here if Milliron catches wind of this posting. *pops popcorn*
Wynn hits the nail on the head with his thought that “organizations aren’t even considering what they want out of Caspio when they sign up.”
I haven’t been to one, but I imagine Caspio’s sessions at IRE and NICAR are slick. I guess a couple of our folks went to them and came back all psyched about using Caspio.
Which brings to my point about Caspio: It’s a great tool for getting simple data online quickly. It’s especially great for taking the load off programmers/designers/car people/whatever, especially if you’re looking to get cheap clicks.
But the idea that Caspio doesn’t require programming just isn’t true. If you want to do anything beyond what comes out of the box, you gotta use programming to adapt it…whether it’s JS for mapping or whatever.
Why not take the time it would take you to learn Caspio’s ins and outs and learn Django/Ruby on Rails?
July 4th, 2008 at 2:53 pm (#)
Good points, Ryan. This stuff isn’t rocket science. It does take an investment, but the payoff is worth it.
Recent examples:
- This http://www.nytimes.com/projects/nyctestscores/ was a day of development from scratch: Me, Rob from the CAR team, and Matt Ericson did an hour or two of front-end touchup. It’s ugly, and kinda not the greatest, but it’s certainly comparable to anything you get out of the box with Caspio.
We are in the process of making a Rails plugin out of this that will allow you to point to a database table, and generate the framework of a search/master/details site. (And not have your data invisible to Google…)
- We’ve turned our Clinton documents site (http://politics.nytimes.com/clinton-schedules) into a Rails plugin. To generate something like this (http://www.nytimes.com/cityroom/documents/koranteng) from a pdf takes a few minutes. Next version will be completely SEO friendly and will include inline, page-level annotation, kind of similar to the way the Django book does comments.
- This (http://www.nytimes.com/interactive/2008/06/02/nyregion/GRAPHIC_CRANE_VIOLATIONS.html) is a really super-simple Google map we threw together on deadline after the crane collapse. Again, not the most compelling thing in the world, but we’re almost done converting this into a basic Rails plugin that will allow you to generate a much better version of this in minutes.
All of these plugins took some initial work, but will give us infinitely more freedom and flexibility than Caspio ever will. And, I won’t be locked into their proprietary software.
For what it’s worth, we’re considering very seriously releasing the document viewer and possibly the other two as open source software, for non-commercial use (sorry, caspio!) if people think they would use them.
July 6th, 2008 at 11:21 pm (#)
[...] smarter people have invested a goodly amount of space to explaining Caspio’s deficiencies, so [...]
July 7th, 2008 at 12:04 am (#)
NYT document viewer looks great. I hadn’t seen that yet.