/[cvs]/joko/doc/topics/gde.txt
ViewVC logotype

Annotation of /joko/doc/topics/gde.txt

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.1 - (hide annotations)
Mon Feb 24 18:21:25 2003 UTC (21 years, 4 months ago) by joko
Branch: MAIN
CVS Tags: HEAD
File MIME type: text/plain
updated

1 joko 1.1 From: "Andreas Motl" <andreas.motl@ilo.de>
2     To: <gpe@handhelds.org>
3     Cc: <brentt@ill-logic.org>,
4     <alan@lxorguk.ukuu.org.uk>,
5     "su" <su@tunemedia.de>,
6     <janosch@ultrajan.de>
7     Subject: Re: SyncML
8     Date: Sat, 22 Feb 2003 20:55:43 +0100
9     MIME-Version: 1.0
10     Content-Type: text/plain;
11     charset="iso-8859-1"
12     Content-Transfer-Encoding: 7bit
13     X-Priority: 3
14     X-MSMail-Priority: Normal
15     X-Mailer: Microsoft Outlook Express 5.50.4807.1700
16     X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
17    
18     Brent Theisen <brentt@ill-logic.org> wrote:
19     > Quoting Alan Cox <alan@lxorguk.ukuu.org.uk>:
20     > > > library/client/server/engine? The reference toolkit released by the
21     > > > SyncML Initiative appears to be licensed
22     > > >(http://www.syncml.org/php/licence.php3)
23     > > > under a BSD style license. If this sort of license is unacceptable,
24     > I have
25     > > I had a dig. Its under a BSD license because you have to license
26     > > patents to implement syncml and do so under as yet undecided terms.
27     > That is unfortunate. The "SyncML Intellectual Property/Patent Issues"
28    
29     I did a short glimpse at syncml.org too and I am concerned about the
30     patent issues as well.
31     The "SyncML Initiative IPR Policy and Declarations" papers
32     at http://www.syncml.org/syncml_ipr.html ff. mention two patents
33     having the following titles:
34     "Synchronization Process Negotiation for Computing Devices"
35     "System and Methods for Synchronizing Information Among
36     Disparate Datasets"
37    
38     [...]
39     > So now what? I don't feel like implementing a brand new synchronization
40     > protocol from scratch and then trying to get application developers to
41     > adopt it.
42     > However if this is what it will take to have a free synchronization
43     > solution
44     > for Linux then so be it, I'll try.
45     > Anybody have an idea how we can solve the synchronization problem?
46    
47     We are currently trying to hack something which aims to address some (all)
48     aspects of a synchronization process. I read the titles from the patent
49     paper
50     mentioned above *after* writing the core and believe it's exactly that what
51     Starfish / syncml.org requires a license for - the "Synchronization Process
52     Negotiation".
53    
54     Our thing is currently implemented using Perl and lives in
55     the Data::Transfer::Sync namespace. It's not (yet) on CPAN!
56     We will release the code under the GPL and/or Perl's Artistic License.
57    
58     Please let me know if you're interested to be informed in detail and/or
59     about future changes/enhancements regarding synchronization with
60     handheld devices. We just try to handle the "heavy stuff" first, that
61     are odbms, rdbms, flat files and structured files (csv, xml and stuff).
62    
63     Already in the queue (and in first pre-alpha stage) is a program called
64     'outlook2ldap' which tries to get address information off m$ outlook by
65     synchronizing MAPI Contact objects against LDAP objects in a generic way
66     handling all the important aspects needed (hierarchical containers, etc.).
67    
68     Any help - of course - is greatly appreciated, this got bigger than expected
69     and the number of "device-handlers" that could be written is countless i
70     believe.
71     That's the main aim of that module: Making it possible to span such a
72     synchronization process across *arbitrary* types of "storages".
73     This is accomplished using the Data::Storage module (also not on CPAN yet)
74     which "just adapts" other storage implementations available from there.
75     Working (syncable=1) handlers are by now: DBI, MAPI, LDAP, Tangram.
76     Directory, File, Files, XML are in preparation, please use rsync for the
77     former three instead. Any ideas for other more important storage-handlers?
78    
79     Pushing the scope of available storage-/device-handlers, we had planned
80     to address some PDAs and/or mobile phones soon.
81     Also interesting would be an attempt to talk to iSync:
82     please visit: http://www.confusticate.com/tech/isync/
83    
84     Another nice feature is that the comparison algorithm itself can be
85     replaced. At the time we have implemented a generic ident/checksum
86     mechanism using md5 digests but an alternative to that would be
87     a more user-friendly algorithm using ident/timestamp pairs.
88    
89     We don't have released anything, the code still lives in the cvs for now
90     which
91     can be visited at http://cvs.netfrag.org/nfo/perl/libs/Data/Transfer/Sync/
92    
93     Please also look at the autogenerated documentation (not very much) at:
94     http://netfrag.org/docs/topics/perl-libs/Data-Transfer-Sync.html
95     http://netfrag.org/docs/topics/perl-libs/Data-Storage.html
96    
97     and maybe (remember it's pre-alpha and most probably broken):
98     http://cvs.netfrag.org/nfo/perl/scripts/outlook2ldap
99    
100    
101    
102     regards, Andreas.
103    
104     p.s.: Do you see any possibility for me to look at some parts of the syncml-
105     specification without breaking anything related to patent stuff and the GPL?
106     I don't do this for now until i'm not completely sure about this, but
107     of course i would like to adapt the interface specs to enable feeding
108     of syncml documents to Data::Transfer::Sync.... Can you help? Please help!
109     Otherwise we would make up a new DTD for this purpose although we
110     currently do very well with 'sync.pl' together with mapping-metadata
111     supplied
112     by an additional Perl-module. (abstraction-level=medium)
113     Please also direct me to other mailing lists if i'm completely out of scope
114     here.
115    

MailToCvsAdmin">MailToCvsAdmin
ViewVC Help
Powered by ViewVC 1.1.26 RSS 2.0 feed