Linked by Thom Holwerda on Mon 10th May 2010 10:03 UTC, submitted by robertson
BeOS & Derivatives Two news items about alternative operating system news in a row? What is this, Christmas? In any case, the Haiku project, the darling of OSNews (hey it's okay now), has released its second alpha release. This new stable development release contains some serious improvements over the first alpha.
Thread beginning with comment 423616
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[2]: Making progresses, but...
by Neolander on Mon 10th May 2010 19:07 UTC in reply to "RE: Making progresses, but..."
Neolander
Member since:
2010-03-08

Unlike Linux, Haiku is designed from the ground up to be a desktop system. Everything from performance tradeoffs and how you install software to presenting a consistent user interface and a stable ABI (less need to compile from source and more friendly for commercial developers) is geared towards being a solid desktop operating system. THAT is what it brings to the table.

That's a fair design goal, but what about the resulting decisions ? I'll answer the following post which goes into more detail about those.

It also has some real niceties that you don't see elsewhere, such as its use of extended attributes, which in the BeOS days was referred to as "a database-like filesystem". Please note that although all major filesystems today support extended attributes, only the BeOS inspired systems really make use of them. The difference lies in the userland. You have to try Haiku and get under its skin to really see what this means. And this part will only become better once the index feeder is in place.

Is this used in order to improve search ? What do I use this feature for ? Which kind of task does it help ?... See where I'm going ? ^^

Edited 2010-05-10 19:08 UTC

Reply Parent Score: 1

StephenBeDoper Member since:
2005-07-06

Is this used in order to improve search ? What do I use this feature for ? Which kind of task does it help ?


In a nutshell, individual filetypes in Haiku are roughly analogous to database tables - and each filetype can be given attributes (roughly analogous to fields in a DB). Then the Tracker (filemanager) lets you display those attributes as columns so you can easily view, edit, or sort by attributes.

The three best-known examples of practical uses are managing contact/address book data, EMail messages, and audio files. So, for example, the filemanager can display columns like track number, artist, title, year, etc for audio files - essentially giving you iTunes-esque functionality, except at the OS level.

So, yes, it does help with search - but there's quite a bit more to it as well. Attributes, queries, etc, are as a fundamental to the way Haiku works as piping and redirection are to the UNIX shell - and, in fact, they're used for many of the same purposes (allowing applications to be small and focused, while still being able to share data & communicate via standardized means).

Reply Parent Score: 2

Neolander Member since:
2010-03-08

In a nutshell, individual filetypes in Haiku are roughly analogous to database tables - and each filetype can be given attributes (roughly analogous to fields in a DB). Then the Tracker (filemanager) lets you display those attributes as columns so you can easily view, edit, or sort by attributes.

The three best-known examples of practical uses are managing contact/address book data, EMail messages, and audio files. So, for example, the filemanager can display columns like track number, artist, title, year, etc for audio files - essentially giving you iTunes-esque functionality, except at the OS level.

So, yes, it does help with search - but there's quite a bit more to it as well. Attributes, queries, etc, are as a fundamental to the way Haiku works as piping and redirection are to the UNIX shell - and, in fact, they're used for many of the same purposes (allowing applications to be small and focused, while still being able to share data & communicate via standardized means).

Okay, if I understand well it is something like OS-managed metadata. Sounds pretty nice, though I wonder how well it would scale if millions of attributes were to appear in various applications with time...

Reply Parent Score: 1