[Red Liner]

I have a predicament. I run a start up design services company and we have a bunch of erratic but artistic designers working with me. My issue is:

1. All of them work on the same client file at different points in time.
2. They all work on their local HDD because working off the server is ridiculously slow when the files in question are upwards of 50mb each
3. I am looking for a solution which will magically update the source file on the server by gathering only the changes made to that file on individual desktops connected to the server.
4. The software will also magically control the number of versions - lets say I want to go back two days and see the file, it should allow me to do that.

This software is similar to what I want but has got horrible reviews on the web. version control software | Adobe Version Cue CS4

Can I have suggestions on best alternatives and prices? Any free ones which are fantastic>?

[SirAlec]

What is you budget.??

[linuxtechie]

Hmm, the files in question, are they text files or compressed ones (binary)?

~LT

[Red Liner]

Originally Posted by SirAlec

What is you budget.??

Lets just say I am looking for the best and then will climb down as I see fit. Since this is basically the lifeblood of my business, I really do not want to cut too many corners.

Originally Posted by linuxtechie

Hmm, the files in question, are they text files or compressed ones (binary)?

~LT

These files are not text files alone, but could include them. These could be adobe files, coral files, basically files designers work on. Hope I answered you right.

[pranavt]

4. The software will also magically control the number of versions - lets say I want to go back two days and see the file, it should allow me to do that.

For version control, subversion is good. But it's for programming/coding, first and foremost. It will store different revisions perfectly in seperate folders and hierarchies as per your needs.

3. I am looking for a solution which will magically update the source file on the server by gathering only the changes made to that file on individual desktops connected to the server.

Merging all the changes together into one image might not be possible I think. A couple of ways would be to extract the changes from the original as a "diff" file and apply a bunch of them together, but that could fail since you're talking about binary files with proprietary formats. Try rsync for something like this, but don't expect any results.


Both the solutions are free and need to be run on the server. Hopefully it's running Linux. If it's not, you'll need to run this in a VM or shell out for a seperate Linux box which will host the subversion server. No idea if I'm correct since I'm not into designing and art files.

[DerAlte]

Subversion (SVN), or even the good old RCS/CVS on Linux, should work well for Adobe / Coral files, most of which are a mix of text (not so apparent, since the file header is a mass of bytes that make sense only in binary) and binary bytes representing pictorial information. The only problem will be making sense of the way differences are reported (the binary info will appear as gibberish). [Aren't SVN, RCS/CVS etc. available as Windows builds too? Check]

Merging branches, as @pranavt pointed out, might cause challenges, since the position of the changes may put the header off kilter.

A slight process change (needs a LOT of discipline in the designers) is to periodically 'freeze' a 'build' (in this case a trusted version of the graphical document) and use it as a template for all further work. All further work should be done in layers that should appear on the template. Deliberately leave empty spaces / pages to accommodate the growing product (ever wonder why some manuals have some pages marked 'this page is intentionally left blank'?). The next build should be created by manually merging the layers to produce a document which will be used as the template for further work.

This 'build' should be created by ONE person only. Everyone else checks in their layer(s) only, even if they have managed to make their own local 'builds' (which can be trashed at any time). Each creative person's output can be a branch, and everyone's output is merged together to produce the 'build'. This way, only the template is the heavy file - all others are smaller files.

@red_liner, you could try out this process even without a Version Control / Configuration Management software. Just make sure everyone sticks to a defined file naming system, i.e. the file name should be composed of Project Id, Template version (or Build No.), Designer Id, layer number, date and version. Sounds complicated, but as the biz owner you should be able to figure out what works best in your environment. You will need a file naming system even when you implement a VC/CM software.

[Red Liner]

Thank you DerAlte. I have done some belated digging on this subversion topic and this is what I managed to get one *expert* to spill out:

_____________________________________
SVN = Subversion
TortoiseSVN = a graphic client for subversion

1download TortoiseSVN (from: subversion: Subversion Packages)
Install, restart computer
Find a place where you want to save your data (make sure there´s enough space in the partition)
Create a directory there, called ¨repos¨ or something. In that directory, right-click, choose ¨TortoiseSVN¨ and click ¨Create Repository Here¨
Go to the folder where all your graphics files areRight-click, go to ¨TortoiseSVN¨, click on ¨Import¨
now you have to point it to the path of the repository say your repository is in a ¨resos¨ directory in ¨E:¨, you would put: file:///E:/repos/as the URL of the repository- then for the ¨Import message¨ give ¨Initial Import" or something... ALWAYS remember to put in messages here otherwise you´ll have hell trying to figure out what changed with each version
Click OK. All your graphic files will enter your repository.
In the same graphic file directory (NOT the repository), add ALL the files there to a ZIP file, copy the ZIP file to some backup location and DELETE everything
now your graphic files directory should be empty. right-click, click on ¨TortoiseSVN Checkout¨
The URL of the repository should automatically be filled in. If its not, give the same file:///E:/repos/ and make sure it says ¨Fully recursive¨ in the dropdown, and click OK

Done!

Here on at the end of every day´s work, right click in the root of your graphic files directory, click ¨Check in¨, type what all you changed and the repository will be updated
6:19 PM sorry, its ¨Commit¨ not check in thats all there is to it and one day if your network cable falls off, or if the client says back to the older version, all you´d do is right-click the file, click revert and choose the old version

so your commit history for files would be
version 1: initial import
version 2: added header
version 3: changed font
it is going to save your PSDs, DOCs, etc PLUS their versions as they change

6:35 PM for now, until all that happens, you could even use dreamhost´s subversion service... thats what i use for all my code. you have an account. just create a subdomain, and create subversion repositories in it via the control panel
and use tortoisesvn as i explained above minus the create repository thingy
make sure the server has lots of free space. its best if you get a machine to handle only this and nothing else

like i said, best is you either get a server just for version control, or alternatively add a huge hard disk to an existing server and create repositories on that

read this for a better explanation: OSS Watch - What is version control? Why is it important for due diligence?
hmm
6:47 PM photoshop has support for it... How to Enable Version Control for Adobe Photoshop Files
6:48 PM he´s explaining using something called filehamster. no clue how good that is but its better than nothing
__________________________________________________ _

Is this possible? What do you guys think? I am keen on implementing this on my Window's server across all kinds of documents.

[null]

SVN is the best for you in my opinion

[katpasin]

I have been using MKS Integrity at my work.

meets all our software configuration requirements.

Not a freeware though

[DerAlte]

Sounds good, @red-liner. of course this is possible. Other than info on what gets stored where deep down, that is all there is to it.

As I had mentioned earlier, you should first start implementing a good disciplined manual process first, and then supplement that with appropriate software.

[Red Liner]

Originally Posted by DerAlte

Sounds good, @red-liner. of course this is possible. Other than info on what gets stored where deep down, that is all there is to it.

As I had mentioned earlier, you should first start implementing a good disciplined manual process first, and then supplement that with appropriate software.

I agree with you DerAlte. Have to sit down with my team and come up with a naming convention and file hierarchy system. Phew!

[skyliner34]

SVN is ideal for this kind of scenarios. When it gets really complicated and if you need the code to be in sync'd geographically then you should consider IBM Rational ClearCase. you can do amazing things with ClearCase. The downside of this tool is cost and maintenance [COLOR=blue][/COLOR] [COLOR=blue][/COLOR]

[Red Liner]

What do you mean *geographically*? Across countries?!

[skyliner34]

Originally Posted by Red Liner

What do you mean *geographically*? Across countries?!

yes, it's between 2 different sites. but that's optional. you buy it only if you need it.
but trust me despite being very expensive it does all we want.