[skarin]

Originally Posted by phanley

I too am writing a OmniSync iPhone connector for Windows. I want to sync with Outlook using .NET. I appreciate the Omni Groups openness with the sync data.

I'm a hardware engineer during the day, so this is a weekend activity for me and it is slow going. But, I have the basics working, getting the XML zip files, extracting them and getting the data into my program. On the other side, I have access to Outlook's tasks and can generate/modify them. For me, now the next challenging part is figuring out how to map the content.xml information to Outlook. My goal is to have two-way sync with the iPhone so that Jello.Dashboard works in Outlook (seamlessly).

--
Paul

I have a similar end-goal. I am currently working on, as a sort of "proof of concept", a "quick add" app, where I can just hotkey it, and add a new item to my Inbox on OmniFocus. That part's pretty easy. I have it creating the .client and .zip with a contents.xml with a new inbox item. Still hung up on getting the directory list from webdav(for the last used tail ID(if that's needed?)), and uploading my client and zip files to the server via webdav. There doesn't seem to be much documentation out there on using webdav from .NET. My ultimate goal is to have it sync with my tasks in outlook. I'm a .NET developer, so if there's any help I can provide on that front please feel free to ask.

[Lizard]

Yes, every zip file should reference the last used tail ID. Otherwise you'll get disconnected transaction errors, and OmniFocus will not incorporate the data from your .zip files.

[skarin]

Thanks for the confirmation on that. Yesterday I created a client and zip file manually and dropped them in my WebDAV folder and synced on my iPhone and it added the tasks in, so now I just need to finish my automation or the upload.

I want to thank you Omni folks again for the pointers and assistance. This kind of customer service is what makes me want to use your products!

[skarin]

Do both tailIds need to be in the client file? The last used and the "new" id?

[Lizard]

20080807012719=opHaPQGlXzC+eg0yGRGfOcd

The first thing is a date string in GMT.
After the = is the first ID.
After the + is the second ID.
If you're writing a file, your first ID should be the same as the second ID of the most recent transaction file you've seen on the server. And you should have a second ID so you or some other client can continue the chain.

[skarin]

20080807012719=eg0yGRGfOcd+ddgsdgwer53.zip
20080807000000=opHaPQGlXzC+eg0yGRGfOcd.zip

In this example the lower file was synced by my device, so I'd take the last part of the name and make that the first part my my new zip, and make up a new unique 11 char string for the 2nd part of my name and that would make the top file, I have that all sorted out I think. The question is, in my client file, do I need both eg0yGRGfOcd and ddgsdgwer53 in the tail id section at the bottom, or just ddgsdgwer53?

[Lizard]

You don't want both of those, only ddgsdgwer53. However, there are cases where you ought to list multiple tail ids.

I'm going to switch to using single letters for the ids, but realize that they're just placeholders for the 11-character ids.

I might have files on the server like so:
A+B
B+C
C+D

Then I just need to list D in the tail ids section.

Or, because of multiple clients syncing, there might also be a file

C+E

In this case, I should list both D and E in the tail ids.

(After the next OmniFocus sync, there might be a file named D+E+F, and then your next .client file should only list F as a tail id.)

[skarin]

Ok, so the tail is basically just letting other clients know where a certain client is synced up through? I just manually synced about 12 tasks into my inbox a while ago and put the two id's in the client file and it worked. Is there any reason where it could mess things up?

[jjlucsy]

Would anyone like to share their code on what they have so far? I'd like to take up the challenge to get something working, but I'd rather not start from scratch unless I have to.
Thanks.

[devn]

1. I would love to know which elements in the XML file are required for an Inbox item. I don't want to use more than I need to. Could you give me the bare requirements here? A short template contents.xml?

2. When I initially create a task manually, should I be using <task op="update" or op="reference"?

3. The question of time is getting me a little bit. Should I literally take every single .zip file in the directory and check for duplicate tail IDs, or should I just be checking the most recent syncs? How integral is the relationship between the .client files and the .zip file? The client file simply acts as a lock and doesn't necessarily mean anything about the .zip file that is created, right?

Thanks,
D