What the heck is Mesh for? People are still talking breathlessly about its potential, but I can never get a clear explanation of how that potential is going to be realized.
Ok, yeah, I understand the file syncing thing. And code syncing. And versioning. Yada yada yada.
But the question is, what is it *really* for. The real world scenario seems to be based on the idea that we are going to want to run code and access data locally on multiple devices. The presumption is that we are going, in the future, to have a phone, and a laptop, and desktop, and perhaps set top box. They will each have their own storage and will want/need to access code and data that is remote, and at times will not have access to the net. And we need a *platform* for this.
The problem is I can only see one really compelling end user relevant usage scenario, and that is moving media around. It makes sense to me that I might want to easily keep my music or video collection in sync across portable devices, or on my TiVo. Beyond that, I can't come up with a single end user scenario that regular folks are going to care about. I keep looking, and every explanation is just too complicated. The descriptions are too long, the words too big, and the concepts too abstract. The master of this discussion by obfuscation is Steve Gillmor:
Does Silverlight intersect with Mesh to produce the Net OS? Answer: Yes. Treadwell calls it orthogonal and complementary. MicroBig language. Can Mesh support Twitter streams orchestrated by identity mapping via affinities and abstracted to devices across OS, mobile, and corporate divides via Silverlight? Yes, but it can do so much more.
I am not the sharpest knife in the drawer, I will admit. But if this is what it takes to describe why Mesh is important...
Fail.
Here's the problem. How many devices is your average person going to have. A phone. A desktop or a laptop, and in rare cases, both. So first, Mesh has to be available on the phone you have. Good luck making that happen with Nokia or Apple, or Google.
I just don't see device or service providing developers (like Apple with iTunes) adopting this in a wide enough fashion to make it useful. Apple won't do it at all, and to the extent anyone else does it will it matter?
The other enemy Microsoft has is time. The target time frame for this has got to be like five to ten years. And regarding things like synchronizing word processing documents between devices, in that time I just think that bandwidth will be reliable enough that people will access stuff just the way they are starting to right now, through a browser. And there will be so many competing ways to do this, like Apple's mobile.me and iTunes and Google Apps with Google Gears that will get right to the point. And people are using or starting to use these kinds of solutions *right now*. No fancy platforms, no run on sentences, just direct easy to understand meat and potatoes.
Honestly, Microsoft CTO Ray Ozzie has been thinking about synchronization for a really long time. This is what Groove, the company Microsoft purchased to get Ozzie, was doing for the preceding five or so years before the acquisition. This singular focus has carried forward to his new job. Unfortunately, sometimes when you have a hammer, and you are intent on using it, everything looks like a nail. The problem is sometimes there just isn’t a nail there so all you get is a bloody thumb. And so for their sakes, I hope Microsoft has a first aid kit at hand and a more compelling plan B.
8 comments:
Have you seen this already?
http://www.joelonsoftware.com/items/2008/05/01.html
: )
Yes, I have been familiar with Joel's Architecture astronauts critique of Ozzie/Groove since he originally made it and his most recent argument. I think my last piece on this may have referenced it, though not sure.
From my experience, the average business app user does want this.
I can't tell you how often I've been in a meeting and someone's using a laptop and says "crap, that presentation is on Bill's computer". And then they spend forever sorting through their email to see if Bill sent them an email with it attached, but because they're on the laptop they have to use OWA to look through their mail and you just want to kill him and yourself for being there. And of course, you just have to say "Is it on the SharePoint server?" And of course nobody has uploaded it to the SharePoint document library because that takes forethought and effort.
So I do see the need there, but to date all the implementations have been just too icky to use.
1. Set windows policies to make a user's document location to be the file server.
FAIL: Only works for Windows machines that log into the corporate network. Doesn't work for Macs or travelling laptops. Also no versioning so files get overwritten.
2. Document Management System (Documentum, SharePoint, whatever)
FAIL: People have to remember to save files to the DMS. (IMO, this is currently still the best solution we have)
3. Groove
FAIL: Because it was a dedicated desktop environment that required setup and consciously using it people still didn't get it.
4. GoogleDocs - AKA just get rid of the desktop apps and store everything online in an online app.
FAIL: Corporate doesn't want its users using a hosted app to edit business documents. Security and interoperability with desktop Office and just the overall paradigm shift just doesn't have traction in corporate.
So I do see the need for some "magic" auto-synchronizing that "just works" and integrates into the standard desktop workflow. But it will also need to support completely transparent synching with various corporate document management servers. I just don't see MS Mesh as being "it" yet.
I've been using it for just my personal items so far. Common images (wallpapers I like, avatars), a few docs (resume for example) and some family photos. I then got my father and brother to sign up and gain access to the family photos folder. The sync is rather nice as long as you don't expect it to show up within a few seconds of it being posted to the cloud and showing up on your local drive. So I do see some value for right now in the day to day experience.
As far as the future, I see it being more as creating ad-hoc work groups on the fly. The trick will be getting finer controls on it while still keeping it relatively simple to use. I'll give it time. I'm not sure where MS is going, but it does seem interesting.
Regarding DMS systems, I think the FAIL is generally how aggravating it is to save to those systems and to use from those systems even once you do remember not to save it to your own desktop.
And, yeah, I think mobile browsing capabilities and hosted services are likely to race (and run ahead of) heavy-handed, over-marketed synch paradigms. Good call on that one.
We have used Groove @work since before MS bought it, and it has always been great for synchronising files between machines, between colleagues. But all the other bits they added -useless. Groove 2007 is the same, only heavier weight. And right now, my mesh beta is even heavier weight than groove is. It brings the vista experience (slow startup, big memory footprint) to WinXP. If and when a version comes out for my windows mobile phone then maybe it will have value.
1. Media sharing is a funny as there is a lot less unique media than there are people (excluding photos). You dont need a private copy of every MP3 you've ever bought and ripped.
2. Mesh seems built on an assumption that you go offline a lot, and synchronizing matters. Once you assume that the network is always there, the rules change. I can't help feeling that Mesh is just trying to keep the old world alive.
Honestly, I like the idea of having my files in a cloud. I use the IDrive backup tool to ship them all out automagically, but I certainly don't use the 'auto-synch' feature because it eats my hard drive alive trying to figure out what to backup and not. I just let it run once at night, when I'm hopefully asleep. If there were a way to do this in 'realtime' without killing my performance, I'd like to give it a try. Groove has not been a good experience for me, nor has any of the other replication technologies I've seen that try to automatically do the work. I like IDrive, and I don't mind the fact that interacting with it is a little more manual. After all, if you have to maintain your own data, you have to ensure: A) you have media that isn't failing. B) If your media does fail, you have a backup, C) if you have a backup, it has to be on media that isn't failing, and also that you have the hardware to read the media, even years down the road. When you outsource your backups, you don't have to cool the servers, maintain the media stores, trasport the media, and/or maintain the old platforms in a locked closet to stand up in case you need to restore something when you've just upgraded everything to your 'new version' of the storage system. In my opinion, outsourcing storage and backup is great, as long as your security situation is worked out properly.
What about this. You go on a vacation with your family. There are multiple people with multiple cameras. How does everyone get all the pictures from the trip? You could make a Mesh folder and share it out to all the people that went and with one drag you could add all the photos you took to the folder.
Post a Comment