OT: Object storage for files

Not specifically related to SED matters, but there are some switched on people here.

We have constant problems with graphics objects, sometimes complex ones, that are embedded into other graphics. The objects are stored as files, the only thing that characterises them is the file name and a standard extension for the type. For example, an object may be entitled Large_round_thing_with_wheels.ext. There's limited opportunity for documenting inside the objects themselves.

That mode of storage does nothing in terms of linking the object with other information that is required to deploy it, and this fact causes huge inefficiencies, as well as creating a great niche for power techos that want to make themselves unsackable. Is anyone aware of a good solution to this problem, the sort of thing I'm envisaging is some sort of storage system where a file object and other data such as documentation are inextricably bound, and easily searchable. Surely something like this exists out there.

Reply to
Bruce Varley
Loading thread data ...

This sounds pretty much like the forks in Mac OS X storage.

--

Tauno Voipio
Reply to
Tauno Voipio

On a sunny day (Tue, 26 Jun 2012 20:27:27 +0800) it happened "Bruce Varley" wrote in :

I am sure it exists 'prefabricated'. but the usual way to do this is using structures in C. My program subtitler is a typical example, it holds many graphics objects, including text, how to display those, where, and when, in a structure for each object.Each object is a separate file, for example big_picture.gif my_movie.mpg etc The structures go into one large file, here called control.ppml Adding a structure element 'xxx_docs.txt (for each object) probably takes 10 minutes coding.

formatting link

If you cannot program in C you are at the mercy of companies who want your money. I am not saying subtitles is the solution for you, just that using 'linked lists' of structures is the normal solution. I have this suspicion that some people call that database, and in fact of course you can use one for that (SQL) up to some point, Overhead, not dynamic perhaps, and of course you also have the file system to 'organize' things if you cannot program, make a separate directory for each object, together with its docs and whatever, it is, in fact, already there. Unix (Linux) has some cool commands like locate to find it in terabyte size storage system in a fraction of a second. A 100 line bash script will display the object and its docs then in a flash. I do not see your problem... Sorry. ;-)

Reply to
Jan Panteltje

You don't really give enough information on the scale of the problem. Are we talking of inventories of a few hundred shapes or a few million?

Developing a grammar to describe the objects concisely in the title and their use syntax in whatever limited commenting is available would be one way forward. Otherwise you need to enforce a must have a corresponding .TXT file rule before allowing it to be used and beat developers over the head with a blunt instrument until they write one.

Sounds like you want a searchable directory of pre existing objects to avoid unnecessary duplication of nominally standard parts.

Text searching these days is childs play on all but the most insane sizes of data.

--
Regards,
Martin Brown
Reply to
Martin Brown

Not knowing your environment, I can't know if these suggestions are worth anything:

1: (kinda suggested already): enforce that "this-n-that.ext" is accompanied by a "thin-n-that.txt" that describes it. You can back this up automagically to some extent by having a batch file or program that sweeps your directories to make sure that the pairing is there -- but it's going to take constant human intervention to make sure that "this-n- that.txt" actually contains useful information. 2: Double check your file format -- if it's a text-based format then there's a good chance that there's room inside the file for comments or other metadata. Depending on your authoring tool, you may or may not be able to guarantee that the data survives an open-edit-save operation, though. 3 - 1001: To repeat: making this work is going to take constant human intervention. Keeping documentation and technical work consistent is a constant struggle, it will be seen as a waste of time by a significant proportion of your players, and so it will take constant exhortation at all levels to keep things current. The reason is that the documentation never matters until later -- the visible result that naturally leads to rewards is the technical work itself, and that can get done without the documentation. It's one of those "the hero that puts out one fire that he was too dumb to prevent gets a medal, the guy who prevents 1000 fires gets an investigation into his expense report".
--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com
Reply to
Tim Wescott

Which were both awesome and completely dis-interoperable. Woohoo!

I wuz gonna say "sounds like a job for tar & gz". Which isn't.

-- Les Cargill

Reply to
Les Cargill

Google for database. I often use databases to store binary blobs and companion information.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------
Reply to
Nico Coesel

Even Windows has a file "properties" area, where you can write arbitrary text and other data associated with a file. I guess there are tools to search the properties.

--
Regards,

Adrian Jansen           adrianjansen at internode dot on dot net
Note reply address is invalid, convert address above to machine form.
Reply to
Adrian Jansen

Roll your own with an Access Database. Actually if you already have a ERP system then, add a field in the part with a link to the object . And if you called (named) it by its part number, that would solve all sorts of linking issues.

Cheers

Reply to
Martin Riddle

"Bruce Varley" wrote in news: snipped-for-privacy@westnet.com.au:

Use GIF files, they have a document block you can fill to your harts desire. Most gif wiewers ignore that block.

Reply to
Sjouke Burry

Only if you grab the wrong one, you get forked..

Reply to
Robert Baer

AFAIK nothing like that exists outside of a relational database, getting your graphics editor to store stuff in the database in a way that is understandable to both the editor(s) and the database is going to be tricky,

--
?? 100% natural
Reply to
Jasen Betts

It can be done. For a price.

AutoCAD supports some of this functionality. It is possible to attach attributes to drawing objects. While AutoCAD itself has no knowledge of what these attributes are, many third party applications (extensions to AutoCAD) can use them to add 'intelligence' to AutoCAD.

However, AutoCAD isn't really an object based CAD environment from the ground up. It is still possible to 'break' your design environment by stepping outside the extension's features and manually adding or tweaking native drawing objects. Example: Consider an AutoCAD app for drawing schematics. Each 'wire' is a line with netlist like attributes that can feed other apps. But then some designer manages to break the design by adding a few native lines to the schematic. They look just like wires, but with none of the necessary attributes. As far as I can tell, its not possible to 'lock down' AutoCAD to prevent this from happening.

If you want an object oriented environment suitable for a certain design environment from the ground up, they are available. But not cheap. And you'll have to be specific about your environment.

--
Paul Hovnanian     mailto:Paul@Hovnanian.com
------------------------------------------------------------------
If God is perfect then why did He create discontinuous functions?
Reply to
Paul Hovnanian P.E.

[thus, how to locate the suckers?]

There's two solutions: either invest in a database that lets you describe the object at great length, and give its (archival-path and/or filename) in the description. MacOS has pretty good search capabilities for any metadata you want to tag onto a file, and there's some open-source efforts: see for more info.

Or, take really good notes in a big text file, like a journal or diary, and use the computer to search that. How about giving yourself a dummy e-mail account and letting Gmail search the entries for you?

Reply to
whit3rd

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.