2009-01-04
Cocoa Python
For a Python programmer Apple Cocoa Programming might seam interesting since Python (and Ruby) next to Objective-C are supported languages from within XCode 3.x. Python and the bridge PyObjC are both part of the developer installation. Therefor they are available out-of-the-box. Also project templates are available within XCode.
There are advantages to program Cocoa through Python. The most interesting might be the transferal of already existing language know-how and Library knowledge. On the other hand there are disadvantages. First XCode supports ObjectiveC clearly better. E.g. autocompletion of your own Objects is not supported for python. Second the mapping of Python Objects to Objective-C Objects ads complexity as described in Epic Intro to PyObjC while the mapping of functions is rather simple. Third the documentation for Cocoa Python is minimal, mostly refering to the Objective-C look-a-like.
Even that the transfer of python knowledge or the ability to enter programming from scratch into a fairly simple and powerfull language makes it desireable. The described issues make a start with Cocoa Python not something for the faint of heart (read: not for the programmer newbie).
There are advantages to program Cocoa through Python. The most interesting might be the transferal of already existing language know-how and Library knowledge. On the other hand there are disadvantages. First XCode supports ObjectiveC clearly better. E.g. autocompletion of your own Objects is not supported for python. Second the mapping of Python Objects to Objective-C Objects ads complexity as described in Epic Intro to PyObjC while the mapping of functions is rather simple. Third the documentation for Cocoa Python is minimal, mostly refering to the Objective-C look-a-like.
Even that the transfer of python knowledge or the ability to enter programming from scratch into a fairly simple and powerfull language makes it desireable. The described issues make a start with Cocoa Python not something for the faint of heart (read: not for the programmer newbie).
2009-01-02
it :: Converging Application Frameworks
There is a convergence of Web Application Frameworks and Application Frameworks in general. Apple is positionened in a unique way to get itself into the center of current development seen at Google, Zoho, Microsoft Live and others.
This research note gives a good insight into what to expect in 2009 and after. Expect this to be a valueable read if you are involved in IT-Strategies, in Telecoms, the Web 2.0, if you engage in Architecture or Development.
This research note gives a good insight into what to expect in 2009 and after. Expect this to be a valueable read if you are involved in IT-Strategies, in Telecoms, the Web 2.0, if you engage in Architecture or Development.
2008-03-02
it :: Office based on Model View Controler
How would Office look if consequently designed in a MVC (Model View Controler) style? What would it mean to the notion of a Office file format at all? What would it mean to shareing?
Let's have a short look into MS Office. The best example is tables. There are some in Powerpoint, Word, Excel, Access. Each controller (user interface) behaves different, the stored data is different (try to cut'n'paste one to the other and you will see). What is even worse is, that the user can not do the same things. Present nicely in Powerpoint but not calculate well, calculate in Excel, but not ensure data quality, ...
Well, the core software would have to retreat to define the major data model. This than can be extended by microformats and other. But the most important would be to have controllers that extend each other E.g. the upper issues in MS Office.
If now the data should be displayed in a very special form, today there is the selection of the respective spreadsheet software. But I think there are good reasons to let this be done by modules. E.g. the list is actually a genealogy. In todays ecosystems there is no way to interprete the table in such a new way. A new software has to be designed, even that 90% looks exactly the same like office.
Let's have a short look into MS Office. The best example is tables. There are some in Powerpoint, Word, Excel, Access. Each controller (user interface) behaves different, the stored data is different (try to cut'n'paste one to the other and you will see). What is even worse is, that the user can not do the same things. Present nicely in Powerpoint but not calculate well, calculate in Excel, but not ensure data quality, ...
Well, the core software would have to retreat to define the major data model. This than can be extended by microformats and other. But the most important would be to have controllers that extend each other E.g. the upper issues in MS Office.
If now the data should be displayed in a very special form, today there is the selection of the respective spreadsheet software. But I think there are good reasons to let this be done by modules. E.g. the list is actually a genealogy. In todays ecosystems there is no way to interprete the table in such a new way. A new software has to be designed, even that 90% looks exactly the same like office.
it :: office tools are so 20th century
You are emprisoned with me to MS Office, Openoffice, Google Apps, Zoho, iWork and the likes -- and you think they don't live up to your expectations? Well, so do I (check also here).
Most work I do in these days I start in MS Office. Mostly it is heavy on Powerpoint, Tables and Excel. Right after some slides I typically wonder why I can't do longer tables, why I can't calculate in them and how I want to include the voicetrack for the ones that have not the right location to hear me presenting the slides.
In other work a Slide grows into an Excel, grows into an MS Access and then last but not least has lost any intelligent way of data presentation.
Well, I think all todays Office options are not up to the task. There are some important things missing out totally: MVC design (model view controler), units in calculations, unsharp data, modular data storage (better than publish/subscribe and ole), built in documentation, multi language, multi output, sharing.
Most work I do in these days I start in MS Office. Mostly it is heavy on Powerpoint, Tables and Excel. Right after some slides I typically wonder why I can't do longer tables, why I can't calculate in them and how I want to include the voicetrack for the ones that have not the right location to hear me presenting the slides.
In other work a Slide grows into an Excel, grows into an MS Access and then last but not least has lost any intelligent way of data presentation.
Well, I think all todays Office options are not up to the task. There are some important things missing out totally: MVC design (model view controler), units in calculations, unsharp data, modular data storage (better than publish/subscribe and ole), built in documentation, multi language, multi output, sharing.
2005-08-04
it :: hate one dimensional hierarchies
Well, if you can't stand stacks of folders nor the resulting one dimensional hierarchies and if the unix-like alias system is not really satisfying your outbreak -- then read on.
Let's get rid of folders and filenames alltogether. People think differently anyhow. Usually a document fits into more than one area. Live is multidimensional anyway. So how may a solution look like?
Operating systems from now on include desktop search engines with full text indexing. Filetypes allow metadata for years. CSV or Subversion allow versioning. Flicker developed tagging into an art form.
What about this: the datastore is always online therefor in the internet at home or centrally hosted. Your content can be in multiple virtual groups at a time. The sharing of this content can be graphically defined through a circle of sharing (radially growing rings from ego to internet with sectors for work, live, family, hobby, vicinity, etc.). If a document or document group is shared to a group represented by a ring-sector this sector will be colored. Separate individuals
are marked by a colored pixel in their ring-sector.
A central issue is easeing read/write access to metadata. On any search with a wrong hit one can easily add a tag to clarify the content of the document. Groups can also be tagged. A strong classification of Metadata ensures richer forms of search. Examples for classification would be EXIF data or versioning date and person stamp of text documents.
Let's get rid of folders and filenames alltogether. People think differently anyhow. Usually a document fits into more than one area. Live is multidimensional anyway. So how may a solution look like?
Operating systems from now on include desktop search engines with full text indexing. Filetypes allow metadata for years. CSV or Subversion allow versioning. Flicker developed tagging into an art form.
What about this: the datastore is always online therefor in the internet at home or centrally hosted. Your content can be in multiple virtual groups at a time. The sharing of this content can be graphically defined through a circle of sharing (radially growing rings from ego to internet with sectors for work, live, family, hobby, vicinity, etc.). If a document or document group is shared to a group represented by a ring-sector this sector will be colored. Separate individuals
are marked by a colored pixel in their ring-sector.
A central issue is easeing read/write access to metadata. On any search with a wrong hit one can easily add a tag to clarify the content of the document. Groups can also be tagged. A strong classification of Metadata ensures richer forms of search. Examples for classification would be EXIF data or versioning date and person stamp of text documents.
ce :: radio of the future
The next generation broadcasting networks (ngBN)are just about to be build around the world. South Korea is currently leading with it's DMB deployment. Now what is in for users?
We've got used to podcasting. RSS is a base technology. The exchange of playlists is a biggy on iTMS and Napster sells subscription based music.
Now I will define the ngBNs as follows. Looking at a region or nation they consist of several DVB-T Multiplex (each 3Mbit/s) and allow virtual channels. Let's take one Multiplex. For mp3 audio this ads up to 24 streams (at 128kbit/s) for 24hours a day.
Now why is this special? Because first of all that adds up to 576 hours of new audio every day on your harddisk-device. Second every music lover in the region/subscription receives the exact same extent of music. At this point the idea of playlists kick in.
What if -- the cultural institution of radio changes from beeing streeming audio to exchanging playlists on the base of available audio-sniplets.
We've got used to podcasting. RSS is a base technology. The exchange of playlists is a biggy on iTMS and Napster sells subscription based music.
Now I will define the ngBNs as follows. Looking at a region or nation they consist of several DVB-T Multiplex (each 3Mbit/s) and allow virtual channels. Let's take one Multiplex. For mp3 audio this ads up to 24 streams (at 128kbit/s) for 24hours a day.
Now why is this special? Because first of all that adds up to 576 hours of new audio every day on your harddisk-device. Second every music lover in the region/subscription receives the exact same extent of music. At this point the idea of playlists kick in.
What if -- the cultural institution of radio changes from beeing streeming audio to exchanging playlists on the base of available audio-sniplets.
2005-08-03
it :: the visual world online
"it :: the visual world" described a method to enhance the satelite globe experience by massive multicreator online efforts.
This method may be used for a special experience by realtime picture or video taking in zones of any special events. RSS is capable to function as syndication tool. Is there a user side interface to link certain picture, audio and video streams comming out of one event a new live experience is created.
Are certain users able to broadcast after linking the outcome the method can complement tv-news.
This method may be used for a special experience by realtime picture or video taking in zones of any special events. RSS is capable to function as syndication tool. Is there a user side interface to link certain picture, audio and video streams comming out of one event a new live experience is created.
Are certain users able to broadcast after linking the outcome the method can complement tv-news.
it :: the visual world
You may wonder what happens next in the vincinity of Google-Keyhole? Well you make it happen. Take pda like devices that are just learning to take medium resolution pictures and talk wifi, edge or umts. Add a GPS and a orientation sensor (tilt, direction).
GPS, orientation data are combined with camera data like angle, focus distance and resolution plus a short description (optional). All that is stored in the exif fields. The picture is sent to a server.
The server maps all received pictures to a virtual globe (best take a satellite imaged globe (like google's earth or MSN's map). Since there may be already some pictures from the same spot the server can now stitch those images together and bring them to a average brightness.
The method works best, if a series is taken from one device at one time.
GPS, orientation data are combined with camera data like angle, focus distance and resolution plus a short description (optional). All that is stored in the exif fields. The picture is sent to a server.
The server maps all received pictures to a virtual globe (best take a satellite imaged globe (like google's earth or MSN's map). Since there may be already some pictures from the same spot the server can now stitch those images together and bring them to a average brightness.
The method works best, if a series is taken from one device at one time.
