2011 status report

[f0dder]

OK, so I haven't done a helluva lot of work on fSekrit since the progress and thoughts threads was started. No lame excuses, just a lot of Real Life^TM

I've decided on Git for version control, as I've verified it can do the history split/merge I want for the "historic & private" version vs. the "new & public" one. Not entirely sure how to wrangle the grafting yet, but I know it can be done, and that's the important thing. I'm not entirely sure when to move things over to Git - or, rather, where to start the public history, as I've already moved the private repository to Git. I could do the remaining cleanup so there's no swear-words or other embarrassing stuff in the codebase and go public there, or I could do the minimum amount of work so there's a working & tested build (which is some effort) until the code base is released. Not entirely sure yet.

So, what's the status right now?

I've pretty much settled on the internal data representation I want for the next version of fSekrit, which will allow for things like multiple tabs in one document and future option expandability without requiring file-format changes. I've updated reader code for v1 and v2 of the fSekrit file format to read into this new internal representation, but I still need to settle on a serialized format of the internal v3 representation, and write load/save code for this. Not terribly complicated, but fairly boring - C# is so much easier than C++, just add [DataContract] and [DataMember] attributes and you're pretty much done . I'm considering whether I should just use Google's protobuf, but on the other hand one of the main selling points of fSekrit is compact size.

And then there's the other stuff from the Progress thread that haven't really been started yet. What I'm currently considering is to finalize "sekritCore", which means verifying that I've made correct flexibility decisions with regards to the v3 document format, finalizing v3 load/save code, and possibly get some unit testing in place for these core features (as far as I can tell, Googles' testing framework is the best bet for C++ code). Once that's done, opensource the project, and start picking away at the ToDo list, one feature at a time.

Any comments, or have people stopped using fSekrit for lack of updates?

[Stoic Joker]

I'm considering whether I should just use Google's protobuf, but on the other hand one of the main selling points of fSekrit is compact size.

-f0dder (March 04, 2011, 01:44 PM)

Stay with C++ - Small and fast is (worth it) good! - Misery loves company...

[cranioscopical]

I'm unqualified to comment on the technical aspects but am delighted to see fSekrit about to move forward.

Thanks for the time and effort that you put into it!

[40hz]

One question: Why open source it? If you're going to be the sole coder and maintainer why bother to go that route with all the aggravation the 'open thing' so often brings with it?
 

[f0dder]

I'm considering whether I should just use Google's protobuf, but on the other hand one of the main selling points of fSekrit is compact size.

-f0dder (March 04, 2011, 01:44 PM)

Stay with C++ - Small and fast is (worth it) good! - Misery loves company...

-Stoic Joker (March 04, 2011, 02:05 PM)

Oh, it's going to stay C++; there's protobuf, protobuf, everywhere

One question: Why open source it? If you're going to be the sole coder and maintainer why bother to go that route with all the aggravation the 'open thing' so often brings with it?
 

-40hz (March 04, 2011, 03:00 PM)

It's a vote of confidence kind of thing. No backdoors, hopefully no nasty bugs, ability to read the documents even when the world has moved from Windows to lunix and can no longer execute the .exe files, the possibility to accept patches from 3rd party developers (especially important since there's some features that plainly don't interest me).

And to ensure the program's life. I'm not going to stick around forever; hopefully for a while yet, but who knows if I might get de-persisted by a drunk bus driver tomorrow?

[Deozaan]

but who knows if I might get de-persisted by a drunk bus driver tomorrow?

-f0dder (March 04, 2011, 03:07 PM)

In that case please release the source today.

[AbteriX]

Any comments, or have people stopped using fSekrit for lack of updates?

-f0dder (March 04, 2011, 01:44 PM)

I like fSekrit as it is right now 
I use it mostly to keep serial numbers for an app in the same [public] folder (or CD) as the setup for this app.
Just double click fSekrit.exe, enter the password, read the serial, close the dialog. Pretty fine.

[kilves76]

Tabs is a great idea to store different profiles/contexts in one file 
  With advent of computing power, maybe consider 1024 bit encryption... maybe combined with "random" encryption variations from user's password or even user selectable (a combination of rol, ror, xor etc mangling stuff, for example pw_letter/mod_3 = use rol/ror/xor).
  Please note that i am not a crypto expert so don't actually know if this would be a security risk instead of an improvement!    But it's an idea that's been in my head for ages, thank you for letting me steam it out.

[f0dder]

With advent of computing power, maybe consider 1024 bit encryption...

-kilves76 (July 04, 2011, 03:37 AM)

Key sizes larger than 256 bits currently tends to mean public/private key based encryption (Like, RSA) rather than symmetric-cipher encryption (like, AES). That's a simplification, but the general idea is that larger bitsize doesn't necessarily mean more secure, especially not if comparing ciphers types that are very different.

maybe combined with "random" encryption variations from user's password or even user selectable (a combination of rol, ror, xor etc mangling stuff, for example pw_letter/mod_3 = use rol/ror/xor).

-kilves76 (July 04, 2011, 03:37 AM)

That sounds fancy, but wouldn't add much real security - I believe in the mathematicians and cryptographers more than the marketing people - and I sincerely belive that if somebody had the capacity to break AES, adding a few ROR and ROLs would only make them LOL

Please note that i am not a crypto expert so don't actually know if this would be a security risk instead of an improvement!    But it's an idea that's been in my head for ages, thank you for letting me steam it out.

-kilves76 (July 04, 2011, 03:37 AM)

Thanks for the ideas - I'm by no means an expert myself... so I let the experts do the hard mathy stuff, and follow their advise

[kilves76]

This just came to me: the ability to create highly random passwords. Choose charset (how large: a-z, with numbers, punctuation, capitalization etc), choose length (even random length between user set min and max), and have it generated. Easy to save to Fsekrit, easy to copypaste to any new website, and let your browser remember it.

[jgpaiva]

kilves76: I'm not sure if fsekrit's intentions are to replace a password manager (but f0dder will surely clarify).
Still, what you mentioned is very similar to the way keepass works.

[f0dder]

I primarily use fSekrit for keeping hold of my various passphrases, but I'm not sure generating passphrases is something I want in core fSekrit - I want to keep it lean_and_mean. And there's definitely enough other tasks I'll have to implement first, anyway.

Hope I can get some vacation time around xmas or January - work has been eating up pretty much all my time & energy.

[wraith808]

work has been eating up pretty much all my time & energy.

-f0dder (November 29, 2011, 10:04 AM)

* wraith808 raises a glass in commiseration. 

[Stoic Joker]

work has been eating up pretty much all my time & energy.

-f0dder (November 29, 2011, 10:04 AM)

* wraith808 raises a glass in commiseration. 

-wraith808 (November 29, 2011, 10:47 AM)

I'll drink to that! ...Christ, I burned out back in July...and I'm still dragging my carcase through the motions.

[wraith808]

Our beta date is next week...



...vodka *does* make everything better.

[Stoic Joker]

[/url])[/i]
...vodka *does* make everything better.

-wraith808 (November 29, 2011, 01:43 PM)