Is Spotlight actually Tunnelvision?
As a web designer with multiple clients and multiple projects, each with a myriad of files to track - it would seem like I would be a great user to take advantage of Spotlight when it debuts with OS X 10.4 Tiger. As I watched the WWDC presentation and how Steve Jobs so easily and fluidly showed how powerful and fast (a point he mentioned at least 20 times) Spotlight was, I admit I was intrigued.
Of particular use is Spotlight integration into System Preferences. Hunting and clicking through preferences is not the best of times. But I only rarely go there, maybe twice a week, when I average 60 hours on the computer. But this is a tiny percentage of my time. it got me wondering if this is just the Google-syndrome that everyone seems to be ailing under. Search is the killer app!
Spotlight in Mail? Umm… okay but don’t we already have a robust search? Spotlight in the Menubar sure is slick, but I can’t imagine using the Spotlight in the Menubar for search for anything other than files. I go to each of the distinct iApps when needing to find pertinent information. Now with the introduction of Widgets, those iApps are much more convenient to access, and offer search that is contextual - where relevancy is more important than 100’s of results for the search term “Paris” when I really wanted all Paris’ that sent me email or Paris related photos. The point is that a user already knows the context of a search query, and will start the query in the relevant Application or information space. Apple can make it easier for the user to get to the appropriate context, the Dashboard/Widgets is a great step.
This meta-data Spotlight idea works great in iTunes. A single type of content, with an incredible amount of meta-data. The important aspect here is that to be truly usable, your content (digital audio files) all need to have the same meta-data fields, and all entered and formatted appropriately. I’ve spent days on making sure my relatively small music collection is useful inside iTunes. The genre entries of “New Age” and “NewAge” are different genres to iTunes and will show up that way. Even worse is Album titles, a slight deviation of a single title will misplace your track forever within iTunes.
And now Apple wants us to manage the meta-data of any content type? Puhlease. I’d rather switch off Spotlight in its entirety rather than be subjected to search results with no chance of being successful, or faced with managing meta-data for every single content type and file on my computer.
I challenge others to figure out a useful query for 70% of mac users, based on meta-data, that is not currently included built in Panther. I can’t think of one.
What I can think of is many ways to improve non-search ways to access content. The Finder, ugh! It is such a failure in so many ways. Why not make the Finder more useful, consistent, predictable, and most importantly fast. I don’t want to search for my stuff, I want to find it!
Spotlight is addressing the market hype of Google, search, and meta-data, while ignoring user’s needs like a proper Finder to find files.
del.icio.us
Digg
Facebook
Slashdot
StumbleUpon
Comments
Don’t be silly! You’d rather go to mail to search for e-mail and rather go to word to search for a word file? Well nothing’s stopping you but I bet the rest of us would enjoy not even having to think about which software we used to compose a document (did I mail that or was it an attachment?) and just get answers. Spotlight will hopefully be what the Mac is all about—the computer gets out of the way!
By the way, smart lists (dynamic finds) look like one of the best features in years (I have items that, based on different meta information about them such as web pages (in the sites folder) and customer’s web site (in the customers folder) go in different places. I am not about to make an alias to everything but a smart list enables me to make a folder called “customer A” and have it dynamically include EVERYTHING for that customer without me having to make copies or aliases or anything. Same is true for mail.
RJZ
Very poor article. If the metadata in iTunes didn’t work then how did you know that “New Age” is different from “Newage” ? The answer is obvious doing a search for “new” or “age” or probably any combination will correctly pull up these files. What you’re complaining about isn’t the results of search or metadata but rather organizational consistency. The search functioned as it should have by pulling up the documents and then it’s up to you to modify them as you see fit.
The point of Spotlight is not to eliminate the usefulness of search in individual applications but to supplement it with a global search capacity that incorporates data from disparate sources. As you know in your web work, a project can easily have parts generated in many applications, and some times you just need to bring up everything; files, correspondence, invoices, etc. to find the pieces across different applications for resolving an issue about a particular project. While expert users, such as yourself, may be able to keep track of whether the desired information is in an email, a Word file, or a Quark document, believe me the average user often needs all the help they can get to track things down. Microsoft and Apple are both investing considerable energy in this area because it is a major usability issue on every platform.
The key for application developers is to expose as much “free” metadata to Spotlight as possible so that “fuzzy” searches may be made such as finding all Photoshop files with five or more Adjustment layers, or Illustrator documents with objects colored in Pantone 239 or iMovie projects that utilize a checkerboard transition. While these may seem like weird queries, any metadata about a file is potentially valuable, especially for graphic files where text may be absent or minimal.
My dream file system is what I call the “big bucket” where everything gets dumped. No folders, no file hierarchy, no endless hours of organization, and a powerful search tool that can find anything everywhere. The Mac originally borrowed the metaphor of files and folders because of its real world familiarity, but the metaphor breaks down badly when information is stored across applications in multiple and unrelated locations. Spotlight transcends the limitations of the Finder and utilizes a simple search approach familiar to most folks who have used a computer. It’s a very, very good thing.
I challenge that the queries in the examples above are for expert users. For novices, a Finder that allows users to organize better is much more useful than global search on meta-data and indexed content.
I think Smart Folders are great, I use them all the time in iTunes. But really, how do you make a smart folder for a client, let’s say its “Apple Inc.”? I have to do several things, all of which support the smart folder functionality which will always be an extra step in my workflow. I could be sure to name all my files with a variation of “Apple Inc.”. And to boot, I’d have to make sure it was properly formatted every time. Because “Apple Inc” is different from “Apple” or from “AppleInc.” Or I could go to every client file and Get Info, and tag the comment as “Apple Inc.”
If you can come up with a non-intrusive way of organizing my Apple Inc. files, that is more powerful than just creating a folder labeled “Apple Inc.” and placing all the files in there, then let me know.
As for the point of search across content types - Let’s say I need Apple Inc. files with email and contacts. Last I heard Microsoft does this with the Office Suite. Not exactly innovative (as S Jobs proclaimed Spotlight is years ahead). I do not deny the usefulness of this type of search. But it is not “fuzzy.” The user still needs to maintain the meta-data to match disparate content types.
The only fuzzy aspect of Spotlight has to be built into an Application - the example shown was System Preferences, and the search for “wifi” pulled up the Airport Prefs. Now this is slick, its contextual and relevant.
But what if I type in “wifi” into the Menubar? It’s too open-ended. It is a recipe for user confusion and failure. Great design sets up users to succeed. Now you could enter “I want to configure my wifi settings” and magically pull up the Airport Prefs. But is this even possible in Spotlight? We shall see, I hope so - and if it is, it will truly be innovative.
I tend to think Nathan is on the right track. As an example, I use launchbar a lot on my Mac. It’s a good experience, lets me use the keyboard more, but it doesn’t radically change my life. My beta ran out, life went on. I’m a little suspicious of speed claims, I can never tell whether steve means “fast on a powermac” or “fast on any mac”. I’m sure that metadata would save me a couple minutes every day on my computer. I’d probably find good uses for it. But would I pay $130 for it and a few other updates which I seem to be fairly content without? Not really. I believe I can speak for average computer users in saying we’re not always interested in a “new and better way”, just a bit more speed and consistency in the old and familiar. Maybe next time, Apple.
hmurchison: I never said meta-data in iTunes doesn’t work. In fact it does, but too strictly. Sure if I did a search for “new” constrained to “Genre” I would no doubt pull up all the correct tracks, no matter how “New Age” was formatted. But I don’t want to do that every time. The built in Browser in iTunes will ALWAYS read the genre “New Age” different than “Newage” and display them in the column browser as such. Ugly, and inelegant: because it requires the user to manually make sure the meta-data is right.
I don’t have an answer for any of this, if I did I sure wouldn’t be writing about it - I’d be designing it. But I think its a worthwhile counter-point of view, and not a “poor article.”
Nathan- Ok I see your reasoning. I thought you were unduly slamming it at first. I see Spotlight as just the beginning. iTunes introducted the concept of searching by metadata limited to one specific genre. Music. Spotlight will extend this to the whole OS via indexing content. However all of this effort seems to be a Trojan Horse for getting consumers acclimated to this type of searching so that the eventual coup de gace can occur and that is a metadata infused filesystem which is what MS Longhorn seeks to have. By implementing it into the fs you have a more elegant system but it puts the onus on develpers to get familiar with the new fs so it’s probably good to do this transformation in steps.
Josh. Thanks to bittorent and other sharing apps non developers are running Tiger now and I’m happy to report that the speed claims are not Steve Jobs RDF. It actually is fast on most systems and that is encouraging. However I would be remiss if I said that Tiger is much more than just Spotlight. It’s a developers bonanza and I expect to see some really nice enhancements to applicatons using new tools in Tiger. Some people may choose to wait before upgrading but there’s no fluff in Tiger. It’s solid stuff.
Interesting how some people are already doubting how effective a powerful technology like Spotlight can be before it’s been given an opportunity to prove itself, for better or worse.
One concern seems to be that Spotlight searches will produce sloppy, open-ended results that reduce its usefulness and create confusion. But how is that any different than current methods of searching that require adjustments and filtering to obtain relevant, meaningful results?
We’re not losing anything with Spotlight but rather gaining a tool that wasn’t available before (often with awkward and frustrating limitations anyway). It doesn’t seem threatening to the “old and familiar”, unless you want it to.
Sure, there are already many applications that offer searching capabilities similar to Spotlight but the lack of integration between them has long been a source of personal frustration. I don’t like my data held hostage to specific applications to gain the benefits that a system-integrated technology like Spotlight can offer.
I certainly share David Chilstom’s “big bucket” vision. I’ve been told that’s sort of like Newton’s Soup, tho’ I have no experience with it.
As David mentioned, some of the physical world metaphors we’ve been using for data organization/representation (e.g. desktop and folders) are changing. Location is becoming increasingly irrelevant. As long as we can find what we’re looking for it doesn’t *need* to matter (as much) where it is, and (for me) it’s a relief not having to care unless I *want* to.
How we choose to organize and represent collections of data culled from various sources can become more a matter of context and convenience, not a hard constraint. The traditional methods will be simply become one of many choices.
Are we steadily moving towards freedom from “monoculture computing” limitations even if we’re still constrained (and at times blinded) by them? Maybe some of it has been stuck at version one for a bit too long?
I’ve kept my devil’s advocate side to all this in check for brevity’s sake… it’s already too long.
Anyway, it was a pleasant surprise discovering an intelligent, provocative discussion about this topic. Thanks!
An update to the conversation: I went off and downloaded Quicksilver, as it seems the closest to the global Spotlight search.
and you know what, its pretty cool. I use it to launch apps, and I love the integration with Address Book. I just wish there was more: integration with all my iApps… and because its so resource intensive - I wish there was different hotkeys to activate search for Apps/iApps vs. search for files.
As a sorta power user I appreciate the time savings… but am still frustrated by the vagueness that I feel is just the inherent downfall of having 100’s of thousands of files to search against.
Of possible interest:
Spotlight on Spotlight
More about Spotlight