Thursday Jan 18, 2007

PIM Sync

Ian Martin's comments on my posting about foxmarks raised enough interesting points for discussion that I felt it merited it's own posting rather than a simple reply to the comments. It also touched on a couple of areas that I had already been planning to blog about. This the whole minefield that is PIM synchronisation.

I wanted to separate my response to Ian's comment into 3 separate but loosely related topics which I believe can be addressed separately. These are:-

  1. What is PIM - does it include bookmarks?
  2. What does "sync" mean?
  3. Is peer to peer the right approach?

 Ian raises the point that should bookmarks be considered part of the PIM suite of information. My first reaction was NO WAY but then I thought about it a bit longer. Certainly in most PIM management solutions today they are NOT considered. Surely the bookmarks are highly dependant on the device that you are using (eg. WAP bookmarks for the phone, HTML & RSS bookmarks for the desktop and xHTML for the PDA). Well yes but then you get into a discussion about point 2 which is what do we mean by "sync" and then my views started to change. So for the moment I am going to park this question saying that yes I believe bookmarks need to become part of the personal information profile - provided that some of the issues over sync that I discuss lower down can be implemented.

So what do we mean when we say "sync". The traditional meaning is that we want a copy of the information that is stored on the desktop to be available on the handheld and that after "sync" any edits made on either device since the last "sync" are duplicated. There are a couple of problems with this approach/definition. Firstly I have more than one "device" and this problem shows no sign of reducing. I have my work desktop, my home desktop, my laptop, my PDA, my iPod and my mobile phone (really pushing the bounds of utilisation and virtualisation ;-) ). I want to keep these in "sync" or at least have the information that I need to hand. This is the point where the word "sync" needs some redefinition. So firstly I want to be able to have control on a per field basis which fields in each record are actually copied to the device. This is available on many sync systems today. (eg. I only want phone numbers on my mobile and not birthdays, addresses, photos, etc). I also want to be able to define which records get copied. For example there are records that I would want to maintain on my central store that I would not want to have on my devices (largely due to relative storage capacity). In an ideal world I would want to be able to have the "sync" implementation manage all of this for me and be able to make available over the wire information that I use infrequently. That is the sync software would load into my mobile the 100 most frequently/recently used numbers (leaving 28 slots free). In the event that a number I was looking for was not stored on the phone it would go out to my networked repository and search there, down loading into the phone if located. Edits and insertions on the phone would be tracked and played back into the central store when a sync occured as well as updating the most commonly used numbers.

By now you should understand my view on peer to peer. I am currently sure that it can't work. There needs to be central "database" of record in which the complete set of information is stored (when I say complete I mean of course all records and all fields). In using a database of record with mutiple devices syncronising at random tracking the "latest version" to include in the database of record has issues. Using peer to peer would be a nightmare.

So back to bookmarks. Provided I could tag them based on their relevance for download into various devices I think that yes they should form part of the PIM profile but until I can do that my bookmarks folder is too large to want to have a copy on my Palm and my phone and I just would not use it.

Now consider the complexity of adding group and family calendars. ;-)

So how do we solve the problem. I firmly believe that this is not for one individual or corporation to solve. The way forward has to be based on open standards for information exchange and sync. In that respect I am firmly in agreement with  Scott Mace. By the way I do not believe that Outlook is the answer. I do not want everything to be tied to my laptop. I need solution that I can use over the air from an internet connection be that LAN, ADSL, WiFi or GPRS. At the moment the most promising contender seems to be SyncML. I plan to do some experiments with it in the furture and will post the results.




« August 2016