+ IRC logs to this effect

[Replies] [Reply] [View by Thread] [Help]
[Back To SNEAK Development Forum]

CinC snarkles

Apr 02 18:59:49 <Cybertnt> snarkles: we should discuss the scope of sneak now that DC and you are here ...
Apr 02 19:01:24 <snarkles> Cybertnt, Ok, by scope to what are you referring?
Apr 02 19:01:44 <Cybertnt> snarkles: ask D-Cypell
Apr 02 19:02:35 <snarkles> lol ok
Apr 02 19:02:44 <snarkles> D-Cypell, Ok, by scope to what are you referring? ;)
Apr 02 19:05:14 <D-Cypell> Sorry, yes, scope...
Apr 02 19:05:32 <D-Cypell> My observations was due to the whole native compliation thingy jiggy
Apr 02 19:05:45 <D-Cypell> Im not a fan of native Java compilation at all, but I can see why it is being proposed
Apr 02 19:06:05 <D-Cypell> essentially it seems that the reason for this proposal is to provide low-end web availability
Apr 02 19:06:17 <D-Cypell> which is not really a Java stronghold
Apr 02 19:06:23 <snarkles> Well.
Apr 02 19:06:26 <snarkles> Here's my thing.
Apr 02 19:06:48 <snarkles> The whole reason SNEAK is as popular as it is now is because it is a web app, and written in a language that every tom, dick, and harry has on his web host.
Apr 02 19:07:18 <snarkles> I can fully see the strengths of picking one language (Java) and writing for multiple platforms, and therefore not maintaining 2 completely different source trees.
Apr 02 19:07:46 <snarkles> However, I worry that by not providing this alternative "Here, just upload these files and you're good to go" functionality of it, it will create a barrier towards the target audience.
Apr 02 19:08:38 <socrat> it HAS to be a one download and ready to go thing imo
Apr 02 19:08:40 <snarkles> The binary compilation thing seems counter-intuitive to me, based on the fact that Java's whole strength is that it's NOT platform-specific, but on the other hand Tomcat is not that common in the low-cost hosting world.
Apr 02 19:08:42 <D-Cypell> Ok im with you... bear with me a sec, im gonna grab a drink and perhaps we can take a stab at this issue
Apr 02 19:08:46 <snarkles> Sure.
Apr 02 19:11:24 * D-Cypell returns
Apr 02 19:12:01 <D-Cypell> It seems the difficulty here is that there isnt a great solution for both the low-end web and the platform independant desktop
Apr 02 19:12:23 <D-Cypell> .net might pull this off eventually but we are not there yet
Apr 02 19:12:48 <snarkles> Right.
Apr 02 19:13:02 <D-Cypell> Infact... if we just got with the requirement 'low-end web' we are just talking perl and php really
Apr 02 19:13:15 <D-Cypell> neither of which make great desktop app languages
Apr 02 19:13:37 <snarkles> Right.
Apr 02 19:13:42 <snarkles> Unless they're interfaces with something like GTK
Apr 02 19:13:45 <snarkles> But that's just ugh. :P
Apr 02 19:14:01 <snarkles> Another alternative is to stick with Java for the desktop version, and make a PHP clone of whatever Java source is there.
Apr 02 19:14:15 <snarkles> Which is not "great" but it solves both problems without compromising either.
Apr 02 19:14:21 <D-Cypell> Something else to consider...
Apr 02 19:14:38 <D-Cypell> something that keeps coming up as an issue during the development of the java app
Apr 02 19:14:41 <D-Cypell> is the encryption types
Apr 02 19:14:51 <socrat> can't you call java methods from php?
Apr 02 19:15:09 <Cybertnt> socrat: if the vm is installed yes
Apr 02 19:15:16 <Cybertnt> but thats experimental and sucks
Apr 02 19:15:20 <socrat> ah :\
Apr 02 19:15:22 <D-Cypell> from a design perspective their seems to be a world of difference between some of them
Apr 02 19:15:24 <snarkles> socrat, There is that functionality but it's considered extremely alpha. It also requires a JVM installed on the host, which is not very commomn tmk
Apr 02 19:15:28 <D-Cypell> There are the 'fun' schemes
Apr 02 19:15:32 <snarkles> Or I could just wait for Cybertnt to respond. ;)
Apr 02 19:15:33 <D-Cypell> like reverse, pig latin etc
Apr 02 19:15:43 <socrat> lol
Apr 02 19:15:57 <D-Cypell> and the more serious ones that make up things like PGP
Apr 02 19:16:08 <D-Cypell> Perhaps this might be a dividing line to consider
Apr 02 19:16:28 <D-Cypell> I would be happy to go to a web app to do a quick cipher, ROT13 or something
Apr 02 19:16:46 <D-Cypell> but my PGP stuff will stay in plain text on my box only
Apr 02 19:16:57 <snarkles> Right.
Apr 02 19:16:59 <D-Cypell> There arent too many people that will want to upload stuff like that to a web app
Apr 02 19:17:00 <Cybertnt> yes .
Apr 02 19:17:16 <Cybertnt> lol ask snarkles what stuff people put in there
Apr 02 19:17:17 <snarkles> I think it would be a farce to offer that kind of functionality over the web anyway.. even if we did use SSL, etc. Too many risks.
Apr 02 19:17:22 <snarkles> lol yeah.
Apr 02 19:17:27 <snarkles> People actually use it for stuff like:
Apr 02 19:17:39 <snarkles> - Generating MD5 password hashes of MySQL databases
Apr 02 19:17:48 <snarkles> - Generating .htpasswd password hashes (DES)
Apr 02 19:18:05 <Cybertnt> if they do MD5 their stupid as mysql has that func built in..
Apr 02 19:18:06 <snarkles> - Transforming a "common" word (such as "password") and using that as their password on sites. :\
Apr 02 19:18:19 <Cybertnt> lol
Apr 02 19:18:20 <snarkles> etc.
Apr 02 19:18:45 <D-Cypell> Ok, thats very interesting
Apr 02 19:20:09 <D-Cypell> I would say that functions like DES would probably be good to include in the desktop release
Apr 02 19:20:13 <snarkles> Then most use it for getting passed dog-simple hacking challenges as well.
Apr 02 19:20:14 <D-Cypell> its easy enough to implement anyway
Apr 02 19:20:16 <snarkles> Yeah
Apr 02 19:20:35 <Cybertnt> so can we define the scope please?
Apr 02 19:20:47 <D-Cypell> Cybertnt: exactly what we are trying to do
Apr 02 19:20:47 * socrat hates scopes
Apr 02 19:21:18 <snarkles> socrat, That's just because of that one "incident" with the Martian and.. oh no, I've said too much. ;)
Apr 02 19:21:19 <Cybertnt> D-Cypell: i know, just trying to make you work faster :p
Apr 02 19:21:25 <D-Cypell> IMHO I would say that perhaps there is an argument for keeping PHP for the web and Java for the desktop
Apr 02 19:21:27 <socrat> lol
Apr 02 19:21:30 <Cybertnt> socrat: tell me :)
Apr 02 19:21:35 <snarkles> Cybertnt, lol patience. :P
Apr 02 19:21:38 * socrat hates the martians
Apr 02 19:21:39 <D-Cypell> but not mirror apps
Apr 02 19:21:44 <snarkles> There's no point in rushing through this conversation. might as well get it right the first time.
Apr 02 19:21:44 <D-Cypell> maybe some overlap
Apr 02 19:21:59 <snarkles> Ok.
Apr 02 19:22:01 <D-Cypell> For example...
Apr 02 19:22:07 <Cybertnt> snarkles: :P i was just joking anyway
Apr 02 19:22:10 <snarkles> Cybertnt, ok. ;)
Apr 02 19:22:23 <D-Cypell> Reverse String = Web, RSA = Desktop, DES = both
Apr 02 19:22:31 <Cybertnt> ...
Apr 02 19:22:39 <Cybertnt> everything in web can be in desk as well
Apr 02 19:22:44 <Cybertnt> web should be a slim version
Apr 02 19:22:46 <Cybertnt> of desk
Apr 02 19:22:55 <socrat> why bother?
Apr 02 19:23:10 <D-Cypell> Cybertnt: that is another option. Personally I am finding some of the schemes difficult to justify in Java and they are complicating the design
Apr 02 19:23:21 <D-Cypell> but it is definatly worth considering that approch
Apr 02 19:23:22 <socrat> even if its not everyone that uses the functions, its still going to be used
Apr 02 19:23:25 <Cybertnt> well our design is perfect for any algo
Apr 02 19:23:47 <D-Cypell> I think you need to be careful about calling any design perfect
Apr 02 19:23:48 <D-Cypell> heh
Apr 02 19:23:50 <D-Cypell> It works
Apr 02 19:24:01 * socrat is a perfect design
Apr 02 19:24:02 <socrat> :P
Apr 02 19:24:08 <D-Cypell> lol socrat
Apr 02 19:24:12 * Cybertnt slaps socrat
Apr 02 19:24:23 <Cybertnt> D-Cypell: its not a perfect design for any problem in the world
Apr 02 19:24:24 <snarkles> lol socrat
Apr 02 19:24:31 <Cybertnt> but it will be perfect for our scope
Apr 02 19:24:32 <Cybertnt> !
Apr 02 19:24:48 <snarkles> D-Cypell, If Pig Latin is your concern, I really don't think anyone will shed tears over its loss in the desktop version. :P
Apr 02 19:25:08 <snarkles> Though it might be fun to put some other stuff in there like morse code, atbash, etc. that also don't translate well to "shift"/"translation" ciphers.
Apr 02 19:25:10 <socrat> ? why would it not be in the destop version?
Apr 02 19:25:38 <snarkles> socrat, Well we've had to make the design of the app more complicated to allow for "oddball" methods like PigLatin.
Apr 02 19:25:45 <socrat> morse code yes yes yes , with a transformation to a .wav that can be downloaded
Apr 02 19:25:47 <snarkles> Although I think in the end the flexibility will pay off.
Apr 02 19:25:57 <snarkles> socrat, I actually saw a site that did that once.
Apr 02 19:26:08 <snarkles> Was pretty cool :)
Apr 02 19:26:09 <socrat> heh nice
Apr 02 19:26:37 <D-Cypell> I guess what i am saying is that there is a fundemental difference between the kind of 'encryption' where you just muck about with. Like reverse, l33t, pig latin or morse code. This is the kind of thing where you are not really using secret data
Apr 02 19:26:52 <D-Cypell> and the kind of encryption that you would use to send emails, cms etc
Apr 02 19:27:06 <D-Cypell> with your personal keyring and strong cryptography functions
Apr 02 19:27:13 <snarkles> Yes, this is true.
Apr 02 19:27:30 <D-Cypell> If we are looking to not implement the exact same thing in both Java and PHP we need to find something to differentiate on
Apr 02 19:27:34 <D-Cypell> perhaps this is one option
Apr 02 19:27:40 <snarkles> I think that's fair.
Apr 02 19:28:11 <snarkles> PGP specifically should not be transported via the web.
Apr 02 19:28:23 <snarkles> Because the whole reason that works is because private keys are kept private.
Apr 02 19:28:35 <D-Cypell> Exactly
Apr 02 19:28:45 <D-Cypell> The other difference is that if I use PGP i will most likely use it often
Apr 02 19:28:57 * socrat fucks the fucking fuckers
Apr 02 19:29:04 <snarkles> But MD5/DES have other uses (passwd generation, etc.) which I think fit well in a 'down and dirty' (oh I just need to quick check something) use (aka: web)
Apr 02 19:29:05 <D-Cypell> where as ROT13 or ASCII2BIN is probably just scratching an itch at the given time
Apr 02 19:29:10 <D-Cypell> Maybe im generalizing there
Apr 02 19:29:12 <snarkles> socrat, LOL
Apr 02 19:29:41 <snarkles> D-Cypell, Well, people who routinely read newsgroups might use ROT13 more often. (although generally clients have this functionality built-in)
Apr 02 19:29:58 <snarkles> And I've seen entire websites that use the ASCII2BIN function. No lie. :P
Apr 02 19:29:59 <D-Cypell> snarks: thats true
Apr 02 19:30:08 <snarkles> People arguing back and forth in binary. *lol*
Apr 02 19:30:17 <snarkles> I should go through my referrer logs and see if it's still there :)
Apr 02 19:30:22 <D-Cypell> heh
Apr 02 19:30:25 <D-Cypell> ok point taken
Apr 02 19:30:34 <snarkles> But I'm not sure where you're going with this.
Apr 02 19:30:50 <snarkles> You're saying, things that will be used often should be in the Desktop version, and not in the Web version?
Apr 02 19:31:02 <D-Cypell> Perhaps nowhere.... I have in my mind this dividing line between the different scheme that I am not sure I can rationalize
Apr 02 19:31:20 <snarkles> Yeah, ok.
Apr 02 19:31:50 <snarkles> Specifically, you think the Desktop version should refrain from "for fun" encryption types, and focus only on the "hardcore" kind, or..?
Apr 02 19:32:12 <D-Cypell> Yes that sounds about right. There may be some overlap... for example...
Apr 02 19:32:32 <snarkles> I think it's a given that the web version should NOT have things that would compromise the user's integrity if used over the web (but PGP is the only example I can think of)
Apr 02 19:32:35 <D-Cypell> If i want the DES of the word 'D-Cypell' it seems like overkill to download an install an application to do DES
Apr 02 19:32:43 <snarkles> Yes.
Apr 02 19:32:52 <D-Cypell> on the other hand... if I want the DES digest of a 100MB file I dont want to upload it to a website
Apr 02 19:32:58 <snarkles> True.
Apr 02 19:33:08 <snarkles> So we could draw the line there.. no file upload capabilities in the web version.
Apr 02 19:33:19 <snarkles> And no "key generation" capabilities.
Apr 02 19:33:30 <D-Cypell> That sounds like a good assertion to me
Apr 02 19:33:31 <snarkles> "If you want these fancy features, please download our desktop version."
Apr 02 19:33:39 <D-Cypell> Yes... exactly
Apr 02 19:33:44 <snarkles> This goes counter to what CyberTNT was proposing though.
Apr 02 19:33:50 <socrat> our "FREE" desktop version
Apr 02 19:33:51 <snarkles> Which was to only maintain a single version of the source.
Apr 02 19:33:54 <Cybertnt> ?
Apr 02 19:34:12 * Cybertnt doesnt care
Apr 02 19:34:20 <snarkles> And then when we do a build, automatically generate a desktop and web version from the same src.
Apr 02 19:34:27 <Cybertnt> i am like half away :/
Apr 02 19:34:36 <snarkles> No, I know. But this is a discussion we've had before. :)
Apr 02 19:34:38 <D-Cypell> I think that is problematic... I like the idea and im a big reuse fan, but Java isnt going to work well on most standard web hosts and PHP isnt gonna work with for desktop apps
Apr 02 19:35:07 <snarkles> Ok, then that's fine. If Java is "not the tool" for simple web apps without a hell of a lot of work, then it's very easy to draw the line.
Apr 02 19:35:40 <snarkles> PHP 5 actually would correspond fairly well with Java.
Apr 02 19:35:47 <snarkles> But unfortunately PHP 5 isn't supported very well. :\
Apr 02 19:35:59 <snarkles> So it'll be some work to get a PHP 4 version of what we have now in terms of class design, etc.
Apr 02 19:36:12 <snarkles> Cos PHP 4's OO capabilities are an absolute disaster. *lol*
Apr 02 19:36:38 <snarkles> The other option is Perl, whcih would be equally, if not more supported than PHP.
Apr 02 19:36:44 <snarkles> And comet keeps pushing for that. ;)
Apr 02 19:37:22 <snarkles> That involves an extra couple steps for the user (noting the path to the perl interpreter, chmoding the file so it's executable, etc.) but might be a way forward.
Apr 02 19:37:24 <socrat> no , perl is a dirty little thing
Apr 02 19:37:30 <snarkles> lol
Apr 02 19:37:35 <D-Cypell> Well it seems that PHP and Perl would compete for the web space. I lean towards PHP because thats what the existing script is
Apr 02 19:37:53 <snarkles> Yeah. It's also a language that's suited specifically for web, which is what we're gearing toward.
Apr 02 19:38:06 <snarkles> And the "upload and it's done" thing is a huge boon to usability.
Apr 02 19:38:25 <D-Cypell> I just had a thought.... if we wanna be really clever...
Apr 02 19:38:30 <D-Cypell> hehehe
Apr 02 19:38:45 <snarkles> Also, in that impromptu skills assessment, PHP was the only language other than Java that almost everyone had some skill in, which means it would be relatively easy to get people to help with it, etc.
Apr 02 19:38:46 <snarkles> Oh yeah? :)
Apr 02 19:38:50 <D-Cypell> The more trivial schemes can be implemented in PHP and placed on the web server
Apr 02 19:38:57 <snarkles> !
Apr 02 19:38:59 <snarkles> Web services? :D
Apr 02 19:39:05 <D-Cypell> YES!!!
Apr 02 19:39:07 <D-Cypell> You beat me to it
Apr 02 19:39:11 <snarkles> LOL Sorry. :)
Apr 02 19:39:14 <D-Cypell> heh
Apr 02 19:39:23 <snarkles> That's kind of creepy. ;) I'll stop that mind-reading thing now. ;)
Apr 02 19:39:33 <D-Cypell> The client can consume a service for the straightforward stuff and you get the reuse
Apr 02 19:39:41 <D-Cypell> and the more complex can be native in the client
Apr 02 19:39:53 <snarkles> And the web service would be something we'd just house at a central place (sneak.cyberarmy.net or whatever)
Apr 02 19:40:06 <snarkles> we wouldn't have to distribute that to clients.. or would we?
Apr 02 19:40:11 <D-Cypell> Sure... can be very simple XML too. SOAP would probably be overkill
Apr 02 19:40:14 <snarkles> I don't know a LOT about web services, have only done some consuming.
Apr 02 19:40:22 <snarkles> Are you familiar with REST?
Apr 02 19:40:33 * D-Cypell shakes head
Apr 02 19:40:58 <snarkles> K, I'm not either. *lol* But it's what Amazon uses as one of their WS methods.. basically, you call a URL with the parameters you want to use and it returns XML based on that.
Apr 02 19:41:00 <snarkles> So like:
Apr 02 19:41:09 <snarkles> xml.amazon.com?isbn=12456557676
Apr 02 19:41:23 <snarkles> Would give you back XML for the book containing title, author,e tc.
Apr 02 19:41:32 <snarkles> (that's simplifying it a bit, but that's the jist of how it works :P)
Apr 02 19:41:40 <D-Cypell> Thats perfect. Thats all that is really needed
Apr 02 19:42:07 <snarkles> Now the only thing about that option is it sort of eliminates the "offline" capabilities of our offline desktop version (for simple stuff).
Apr 02 19:42:23 <snarkles> Can I just ask at this point why we don't want to have the simple things in our Java version?
Apr 02 19:42:40 <D-Cypell> It would be possible to implement the standard web screens using the same functionality and use XSLT to get the 'prettys'
Apr 02 19:43:28 <D-Cypell> snarks: Thats true. We need to consider how big an issue that is.
Apr 02 19:44:01 <snarkles> Well, what is the issue exactly..?
Apr 02 19:44:03 <D-Cypell> snarks: Well, i dont know that we dont
Apr 02 19:44:21 <D-Cypell> snarks: sorry that was in answer to the 'no more 100% offline'
Apr 02 19:44:28 <snarkles> Basically that we're essentially treating all algorithms as "equal" under this design?
Apr 02 19:44:38 <snarkles> Yeah, got it.
Apr 02 19:46:04 <D-Cypell> Well... we certainly can implement all algorithms in both places. Its extra work, and it would probably be a good idea to set priorities for the first release
Apr 02 19:46:16 <D-Cypell> Maybe do 4 different strong crypto functions end to end
Apr 02 19:46:29 <D-Cypell> and then build out horizontally
Apr 02 19:46:35 <snarkles> OK, fair enough.
Apr 02 19:46:51 <snarkles> Pig latin, l33tsp34k, I could probably whip off right now in about 10 minutes though,.
Apr 02 19:47:00 <snarkles> So I don't see them as being a lot of extra work,
Apr 02 19:47:06 <snarkles> unless I'm missing something though.
Apr 02 19:47:38 <snarkles> The couple I'm not quite sure how to pull off is URLEncode and HTMLEntities, since that's taking a single character and converting it to multiple.
Apr 02 19:47:52 <snarkles> But I definitely don't want to lose those because I use them all the time for random things.
Apr 02 19:48:01 <snarkles> (encrypting e-mail addresses, etc.)
Apr 02 19:48:45 <snarkles> Or are you referring to the fact that the design is somewhat "generalized" so that it will encompass everything from Pig Latin to Blowfish?
Apr 02 19:49:35 <D-Cypell> I wonder whether it is a little over generalized, and over generalization can be a killer when you get towards the end of projects
Apr 02 19:50:11 <D-Cypell> It is better to do 2 or 3 things than to not do 50 :o)
Apr 02 19:50:23 <D-Cypell> but i dont know
Apr 02 19:50:38 <D-Cypell> Its hard to tell because I have not really dont this kind of thing before
Apr 02 19:50:54 <snarkles> Yeah I think we're all in the same boat there. ;)
Apr 02 19:50:58 <snarkles> Personally, though, I like the design.
Apr 02 19:51:03 <D-Cypell> I wrote a PGP client once, just for kicks, but ive not really attempted a large scale cryptography framework
Apr 02 19:51:09 <snarkles> Cos if there's one thing that changes constantly it's what crypt types SNEAK handles
Apr 02 19:51:24 <snarkles> It started out just being base64 and URLEncode, because I was doing like Try2Hack at the time.
Apr 02 19:51:33 <snarkles> Then I ran into another one which made me do something else, so I added that.
Apr 02 19:51:54 <snarkles> Then I ran into a situation where I needed to create a .htpasswd file and had no shell access to run htpasswd command. So I created that functionality. etc.
Apr 02 19:52:28 <snarkles> So by the design being generalized like that, it makes it so I can add anything that comes up very easily, by just implenenting a couple interfaces, and I'm done.
Apr 02 19:52:41 <D-Cypell> Agreed
Apr 02 19:52:52 <D-Cypell> Thats cool with me
Apr 02 19:52:56 <snarkles> As opposed to the version now, where I have to change 1. the drop down, 2. the function list, 3. the big case statement that decides which function to call, etc.
Apr 02 19:53:47 <snarkles> But I do kind of see what you're saying.
Apr 02 19:54:04 <snarkles> I thought the more "low-level" classes like Cipher and Hash were to help differntiate the different encryption types tho?
Apr 02 19:55:02 <snarkles> Or are you referring even more to the "big picture" scope of this thing eventually handling file/network encryption, etc.?
Apr 02 19:56:11 <D-Cypell> I dont really know. It so hard to rationalize this stuff because its just based on past experience and had become more a gut feeling than anything else.
Apr 02 19:56:23 <snarkles> Heh, yeah I hear ya. :)
Apr 02 19:56:31 <D-Cypell> The thing with generalization is that everything you 'allow for' has to be 'considered for' in the future.
Apr 02 19:56:39 <snarkles> But basically your concern is about generalization of the design, yes?
Apr 02 19:56:52 <snarkles> Hm. Could you explain?
Apr 02 19:57:00 <D-Cypell> We we design the gui.... we have to be always thinking "How will this work for....X Y or Z"
Apr 02 19:57:06 <D-Cypell> That is the cost
Apr 02 19:57:09 <snarkles> Ah, ok.
Apr 02 19:57:16 <snarkles> Yeah, that makes sense.
Apr 02 19:57:24 <D-Cypell> The flexibility of generalization doesnt ever come for free
Apr 02 19:57:38 <snarkles> This is how I pictured the development of this app. And you can tell me right now if I'm completely offbase and such (better to find out now ;))
Apr 02 19:57:40 <D-Cypell> The XP school of thought is to do what you NEED today and worry about what might happen when it does
Apr 02 19:58:13 <D-Cypell> snarks: no... there is never a right or wrong answer to this. There are definate merits in the way it is currently designed
Apr 02 19:58:25 <snarkles> Version 1.0: More or less a carbon-copy of the current SNEAK script that does everything the current version does, only without the requirement for a web connection
Apr 02 19:58:36 <snarkles> Version 2.0: Implementation of higher-level and additional crypto tools
Apr 02 19:58:44 <snarkles> Version 3.0: Ability to handle files/networking/etc.
Apr 02 19:58:48 <snarkles> Version 4.0: Whatever.
Apr 02 19:58:54 <D-Cypell> Ok thats good because that sets a good roadmap
Apr 02 19:59:08 <snarkles> I think you're right re: your comment about XP.
Apr 02 19:59:13 <snarkles> That's the way I've been taught to do things as well.
Apr 02 19:59:47 <snarkles> I think that we just had the advantage of having somewhat of a "roadmap" and altered the design that way to start.
Apr 02 19:59:57 <snarkles> Which I don't see as a "bad" thing, although it's probably unconventional.
Apr 02 20:00:15 <D-Cypell> The problem with the XP approch though is it doesnt always mesh that well with some time honoured OOP concepts. Its natural to abstract and generalize in OOP. So the challenge for the OOP XP programmer is to find the line. And that will be different for everyone
Apr 02 20:00:26 <snarkles> Yeah, true.
Apr 02 20:00:45 <-- Cybertnt has quit (Ping timeout)
Apr 02 20:01:02 <snarkles> If this was your XP project at work, how would you guys have handled it? Probably just a simple "CryptMethod" class and all the various algorithms inheriting directly from that, or..?
Apr 02 20:01:35 <snarkles> And then as the need for other crypt tools (say, PGP key generation, network file stream encryption, etc.) came about, started abstracting the design, including more interfaces and stuff etc?
Apr 02 20:01:56 <D-Cypell> That is a difficult question to answer because usually what happens is exactly what we are doing now. Each person throws in their opinion and we wittle until we get a good answer.
Apr 02 20:02:08 <snarkles> Since I've only done team development in school, I'm honestly pretty ignorant when it comes to this kind of stuff.
Apr 02 20:02:22 <snarkles> Ah, well good to know we're not that far off the mark. ;)
Apr 02 20:02:30 <D-Cypell> Not at all :o)
Apr 02 20:02:42 <D-Cypell> Its just a little slower in text
Apr 02 20:02:47 <D-Cypell> and less black eyes ;o)
Apr 02 20:02:50 <snarkles> lol yeah. And without whiteboards, etc. ;)
Apr 02 20:02:51 <snarkles> LOL!
Apr 02 20:03:18 <snarkles> How about when this project is "done" we start on a collaborative development tool that lets you punch each other in the eye? ;)
Apr 02 20:03:40 <D-Cypell> heh a new set of MIME types
Apr 02 20:03:46 <D-Cypell> assault/punch
Apr 02 20:03:49 <snarkles> LOL!!
Apr 02 20:03:51 <D-Cypell> assault/kick
Apr 02 20:03:56 * D-Cypell chuckles
Apr 02 20:04:01 <snarkles> assault/britney_spears
Apr 02 20:04:08 <D-Cypell> LMAO!!
Apr 02 20:04:15 <snarkles> (reserved only for the most violent confrontations ;))
Apr 02 20:04:55 <snarkles> Your other concern too was that it's more easy to abstract a design that's too specific than "specify" a design that's too abstract, yes?
Apr 02 20:06:02 <snarkles> brb one sec
Apr 02 20:06:08 <D-Cypell> Ok
Apr 02 20:06:53 <D-Cypell> But yes... absolutely. My IDE has many refactoring tools. One is 'extract interface/superclass'..... One is not 'Consolidate hierarchy'
Apr 02 20:08:51 <snarkles> Ha, indeed. :)
Apr 02 20:09:21 <D-Cypell> I think that replicating the functionality of the current sneak script for V1 is a great idea
Apr 02 20:09:29 <D-Cypell> that sets some nice, clear boundaries
Apr 02 20:09:56 <snarkles> Ok, cool.
Apr 02 20:10:11 <D-Cypell> I also think that the current design works quite well for that
Apr 02 20:10:25 <snarkles> Version 2.0 is still a little fuzzy for me, but I'd like to maybe take a round of peoples' suggestions and go from there.
Apr 02 20:10:34 <snarkles> And it's a moot point until we get 1.0 under our belts anyway. ;P
Apr 02 20:10:35 <D-Cypell> We can probably drop the AlgorithmProperty stuff too
Apr 02 20:10:53 <snarkles> k one sec, I have to look at this again cos I forget. *lol*
Apr 02 20:12:43 <snarkles> K, that's basically for sending in parameters such as "salt" to DES/MD5 functions, or..?
Apr 02 20:12:54 <snarkles> (I mean methods when I say functions ignore me and my procedural ways :P)
Apr 02 20:14:22 <D-Cypell> Yes, the idea was that it could be used to provide 'extra' configuration info for an algorithm. A salt would be one thing, a RSA key another
Apr 02 20:14:24 <D-Cypell> for example
Apr 02 20:14:58 <D-Cypell> I guess there is no need to remove it
Apr 02 20:15:11 <D-Cypell> but it probably wont be used too much in v1
Apr 02 20:15:16 <snarkles> Ok.
Apr 02 20:15:24 <D-Cypell> at least the existing script doesnt need it
Apr 02 20:15:31 <snarkles> I'd prefer to leave it if possibl.e
Apr 02 20:15:33 <snarkles> possible*
Apr 02 20:15:42 <snarkles> Because otherwise we have to go and change the damn class diagram again. *lol* ;)
Apr 02 20:16:02 <D-Cypell> heh ok
Apr 02 20:16:12 <snarkles> (I know that's a stupid reason for keeping something, but if it's not hurting anything, then.. :))
Apr 02 20:16:19 <D-Cypell> Ok thats great. So we have a nice clear scope
Apr 02 20:16:24 <snarkles> Ok cool.
Apr 02 20:16:34 <snarkles> Are you ok with that scope?
Apr 02 20:16:39 <snarkles> i.e. does it make sense, etc.?
Apr 02 20:16:45 <D-Cypell> I think its excellent
Apr 02 20:16:51 <snarkles> Oh, spiffy! *lol*
Apr 02 20:17:07 <snarkles> I'll make a post to the effect of "development roadmap" to get everyone on the same page.
Apr 02 20:17:33 <snarkles> And layout 1.0, 2.0, 3.0...
Apr 02 20:17:40 <snarkles> ANything else you think I should be covering?
Apr 02 20:18:33 <snarkles> I'll also throw the logs there for the folks who missed it (cybertnt, saj, etc.)
Apr 02 20:18:35 <D-Cypell> I dont think so.
Apr 02 20:18:45 <D-Cypell> Ok great

Replies:

Development Roadmap - CinC snarkles 2005-04-05 14:19:17
- >+ IRC logs to this effect - CinC snarkles 2005-04-05 14:20:46 <

Guest:
Subject:
Message:	On 2005-04-05 14:20:46, snarkles wrote >Apr 02 18:59:49 <Cybertnt> snarkles: we should discuss the scope of sneak now that DC and you are here ... >Apr 02 19:01:24 <snarkles> Cybertnt, Ok, by scope to what are you referring? >Apr 02 19:01:44 <Cybertnt> snarkles: ask D-Cypell >Apr 02 19:02:35 <snarkles> lol ok >Apr 02 19:02:44 <snarkles> D-Cypell, Ok, by scope to what are you referring? ;) >Apr 02 19:05:14 <D-Cypell> Sorry, yes, scope... >Apr 02 19:05:32 <D-Cypell> My observations was due to the whole native compliation thingy jiggy >Apr 02 19:05:45 <D-Cypell> Im not a fan of native Java compilation at all, but I can see why it is being proposed >Apr 02 19:06:05 <D-Cypell> essentially it seems that the reason for this proposal is to provide low-end web availability >Apr 02 19:06:17 <D-Cypell> which is not really a Java stronghold >Apr 02 19:06:23 <snarkles> Well. >Apr 02 19:06:26 <snarkles> Here's my thing. >Apr 02 19:06:48 <snarkles> The whole reason SNEAK is as popular as it is now is because it is a web app, and written in a language that every tom, dick, and harry has on his web host. >Apr 02 19:07:18 <snarkles> I can fully see the strengths of picking one language (Java) and writing for multiple platforms, and therefore not maintaining 2 completely different source trees. >Apr 02 19:07:46 <snarkles> However, I worry that by not providing this alternative "Here, just upload these files and you're good to go" functionality of it, it will create a barrier towards the target audience. >Apr 02 19:08:38 <socrat> it HAS to be a one download and ready to go thing imo >Apr 02 19:08:40 <snarkles> The binary compilation thing seems counter-intuitive to me, based on the fact that Java's whole strength is that it's NOT platform-specific, but on the other hand Tomcat is not that common in the low-cost hosting world. >Apr 02 19:08:42 <D-Cypell> Ok im with you... bear with me a sec, im gonna grab a drink and perhaps we can take a stab at this issue >Apr 02 19:08:46 <snarkles> Sure. >Apr 02 19:11:24 * D-Cypell returns >Apr 02 19:12:01 <D-Cypell> It seems the difficulty here is that there isnt a great solution for both the low-end web and the platform independant desktop >Apr 02 19:12:23 <D-Cypell> .net might pull this off eventually but we are not there yet >Apr 02 19:12:48 <snarkles> Right. >Apr 02 19:13:02 <D-Cypell> Infact... if we just got with the requirement 'low-end web' we are just talking perl and php really >Apr 02 19:13:15 <D-Cypell> neither of which make great desktop app languages >Apr 02 19:13:37 <snarkles> Right. >Apr 02 19:13:42 <snarkles> Unless they're interfaces with something like GTK >Apr 02 19:13:45 <snarkles> But that's just ugh. :P >Apr 02 19:14:01 <snarkles> Another alternative is to stick with Java for the desktop version, and make a PHP clone of whatever Java source is there. >Apr 02 19:14:15 <snarkles> Which is not "great" but it solves both problems without compromising either. >Apr 02 19:14:21 <D-Cypell> Something else to consider... >Apr 02 19:14:38 <D-Cypell> something that keeps coming up as an issue during the development of the java app >Apr 02 19:14:41 <D-Cypell> is the encryption types >Apr 02 19:14:51 <socrat> can't you call java methods from php? >Apr 02 19:15:09 <Cybertnt> socrat: if the vm is installed yes >Apr 02 19:15:16 <Cybertnt> but thats experimental and sucks >Apr 02 19:15:20 <socrat> ah :\ >Apr 02 19:15:22 <D-Cypell> from a design perspective their seems to be a world of difference between some of them >Apr 02 19:15:24 <snarkles> socrat, There is that functionality but it's considered extremely alpha. It also requires a JVM installed on the host, which is not very commomn tmk >Apr 02 19:15:28 <D-Cypell> There are the 'fun' schemes >Apr 02 19:15:32 <snarkles> Or I could just wait for Cybertnt to respond. ;) >Apr 02 19:15:33 <D-Cypell> like reverse, pig latin etc >Apr 02 19:15:43 <socrat> lol >Apr 02 19:15:57 <D-Cypell> and the more serious ones that make up things like PGP >Apr 02 19:16:08 <D-Cypell> Perhaps this might be a dividing line to consider >Apr 02 19:16:28 <D-Cypell> I would be happy to go to a web app to do a quick cipher, ROT13 or something >Apr 02 19:16:46 <D-Cypell> but my PGP stuff will stay in plain text on my box only >Apr 02 19:16:57 <snarkles> Right. >Apr 02 19:16:59 <D-Cypell> There arent too many people that will want to upload stuff like that to a web app >Apr 02 19:17:00 <Cybertnt> yes . >Apr 02 19:17:16 <Cybertnt> lol ask snarkles what stuff people put in there >Apr 02 19:17:17 <snarkles> I think it would be a farce to offer that kind of functionality over the web anyway.. even if we did use SSL, etc. Too many risks. >Apr 02 19:17:22 <snarkles> lol yeah. >Apr 02 19:17:27 <snarkles> People actually use it for stuff like: >Apr 02 19:17:39 <snarkles> - Generating MD5 password hashes of MySQL databases >Apr 02 19:17:48 <snarkles> - Generating .htpasswd password hashes (DES) >Apr 02 19:18:05 <Cybertnt> if they do MD5 their stupid as mysql has that func built in.. >Apr 02 19:18:06 <snarkles> - Transforming a "common" word (such as "password") and using that as their password on sites. :\ >Apr 02 19:18:19 <Cybertnt> lol >Apr 02 19:18:20 <snarkles> etc. >Apr 02 19:18:45 <D-Cypell> Ok, thats very interesting >Apr 02 19:20:09 <D-Cypell> I would say that functions like DES would probably be good to include in the desktop release >Apr 02 19:20:13 <snarkles> Then most use it for getting passed dog-simple hacking challenges as well. >Apr 02 19:20:14 <D-Cypell> its easy enough to implement anyway >Apr 02 19:20:16 <snarkles> Yeah >Apr 02 19:20:35 <Cybertnt> so can we define the scope please? >Apr 02 19:20:47 <D-Cypell> Cybertnt: exactly what we are trying to do >Apr 02 19:20:47 * socrat hates scopes >Apr 02 19:21:18 <snarkles> socrat, That's just because of that one "incident" with the Martian and.. oh no, I've said too much. ;) >Apr 02 19:21:19 <Cybertnt> D-Cypell: i know, just trying to make you work faster :p >Apr 02 19:21:25 <D-Cypell> IMHO I would say that perhaps there is an argument for keeping PHP for the web and Java for the desktop >Apr 02 19:21:27 <socrat> lol >Apr 02 19:21:30 <Cybertnt> socrat: tell me :) >Apr 02 19:21:35 <snarkles> Cybertnt, lol patience. :P >Apr 02 19:21:38 * socrat hates the martians >Apr 02 19:21:39 <D-Cypell> but not mirror apps >Apr 02 19:21:44 <snarkles> There's no point in rushing through this conversation. might as well get it right the first time. >Apr 02 19:21:44 <D-Cypell> maybe some overlap >Apr 02 19:21:59 <snarkles> Ok. >Apr 02 19:22:01 <D-Cypell> For example... >Apr 02 19:22:07 <Cybertnt> snarkles: :P i was just joking anyway >Apr 02 19:22:10 <snarkles> Cybertnt, ok. ;) >Apr 02 19:22:23 <D-Cypell> Reverse String = Web, RSA = Desktop, DES = both >Apr 02 19:22:31 <Cybertnt> ... >Apr 02 19:22:39 <Cybertnt> everything in web can be in desk as well >Apr 02 19:22:44 <Cybertnt> web should be a slim version >Apr 02 19:22:46 <Cybertnt> of desk >Apr 02 19:22:55 <socrat> why bother? >Apr 02 19:23:10 <D-Cypell> Cybertnt: that is another option. Personally I am finding some of the schemes difficult to justify in Java and they are complicating the design >Apr 02 19:23:21 <D-Cypell> but it is definatly worth considering that approch >Apr 02 19:23:22 <socrat> even if its not everyone that uses the functions, its still going to be used >Apr 02 19:23:25 <Cybertnt> well our design is perfect for any algo >Apr 02 19:23:47 <D-Cypell> I think you need to be careful about calling any design perfect >Apr 02 19:23:48 <D-Cypell> heh >Apr 02 19:23:50 <D-Cypell> It works >Apr 02 19:24:01 * socrat is a perfect design >Apr 02 19:24:02 <socrat> :P >Apr 02 19:24:08 <D-Cypell> lol socrat >Apr 02 19:24:12 * Cybertnt slaps socrat >Apr 02 19:24:23 <Cybertnt> D-Cypell: its not a perfect design for any problem in the world >Apr 02 19:24:24 <snarkles> lol socrat >Apr 02 19:24:31 <Cybertnt> but it will be perfect for our scope >Apr 02 19:24:32 <Cybertnt> ! >Apr 02 19:24:48 <snarkles> D-Cypell, If Pig Latin is your concern, I really don't think anyone will shed tears over its loss in the desktop version. :P >Apr 02 19:25:08 <snarkles> Though it might be fun to put some other stuff in there like morse code, atbash, etc. that also don't translate well to "shift"/"translation" ciphers. >Apr 02 19:25:10 <socrat> ? why would it not be in the destop version? >Apr 02 19:25:38 <snarkles> socrat, Well we've had to make the design of the app more complicated to allow for "oddball" methods like PigLatin. >Apr 02 19:25:45 <socrat> morse code yes yes yes , with a transformation to a .wav that can be downloaded >Apr 02 19:25:47 <snarkles> Although I think in the end the flexibility will pay off. >Apr 02 19:25:57 <snarkles> socrat, I actually saw a site that did that once. >Apr 02 19:26:08 <snarkles> Was pretty cool :) >Apr 02 19:26:09 <socrat> heh nice >Apr 02 19:26:37 <D-Cypell> I guess what i am saying is that there is a fundemental difference between the kind of 'encryption' where you just muck about with. Like reverse, l33t, pig latin or morse code. This is the kind of thing where you are not really using secret data >Apr 02 19:26:52 <D-Cypell> and the kind of encryption that you would use to send emails, cms etc >Apr 02 19:27:06 <D-Cypell> with your personal keyring and strong cryptography functions >Apr 02 19:27:13 <snarkles> Yes, this is true. >Apr 02 19:27:30 <D-Cypell> If we are looking to not implement the exact same thing in both Java and PHP we need to find something to differentiate on >Apr 02 19:27:34 <D-Cypell> perhaps this is one option >Apr 02 19:27:40 <snarkles> I think that's fair. >Apr 02 19:28:11 <snarkles> PGP specifically should not be transported via the web. >Apr 02 19:28:23 <snarkles> Because the whole reason that works is because private keys are kept private. >Apr 02 19:28:35 <D-Cypell> Exactly >Apr 02 19:28:45 <D-Cypell> The other difference is that if I use PGP i will most likely use it often >Apr 02 19:28:57 * socrat fucks the fucking fuckers >Apr 02 19:29:04 <snarkles> But MD5/DES have other uses (passwd generation, etc.) which I think fit well in a 'down and dirty' (oh I just need to quick check something) use (aka: web) >Apr 02 19:29:05 <D-Cypell> where as ROT13 or ASCII2BIN is probably just scratching an itch at the given time >Apr 02 19:29:10 <D-Cypell> Maybe im generalizing there >Apr 02 19:29:12 <snarkles> socrat, LOL >Apr 02 19:29:41 <snarkles> D-Cypell, Well, people who routinely read newsgroups might use ROT13 more often. (although generally clients have this functionality built-in) >Apr 02 19:29:58 <snarkles> And I've seen entire websites that use the ASCII2BIN function. No lie. :P >Apr 02 19:29:59 <D-Cypell> snarks: thats true >Apr 02 19:30:08 <snarkles> People arguing back and forth in binary. lol >Apr 02 19:30:17 <snarkles> I should go through my referrer logs and see if it's still there :) >Apr 02 19:30:22 <D-Cypell> heh >Apr 02 19:30:25 <D-Cypell> ok point taken >Apr 02 19:30:34 <snarkles> But I'm not sure where you're going with this. >Apr 02 19:30:50 <snarkles> You're saying, things that will be used often should be in the Desktop version, and not in the Web version? >Apr 02 19:31:02 <D-Cypell> Perhaps nowhere.... I have in my mind this dividing line between the different scheme that I am not sure I can rationalize >Apr 02 19:31:20 <snarkles> Yeah, ok. >Apr 02 19:31:50 <snarkles> Specifically, you think the Desktop version should refrain from "for fun" encryption types, and focus only on the "hardcore" kind, or..? >Apr 02 19:32:12 <D-Cypell> Yes that sounds about right. There may be some overlap... for example... >Apr 02 19:32:32 <snarkles> I think it's a given that the web version should NOT have things that would compromise the user's integrity if used over the web (but PGP is the only example I can think of) >Apr 02 19:32:35 <D-Cypell> If i want the DES of the word 'D-Cypell' it seems like overkill to download an install an application to do DES >Apr 02 19:32:43 <snarkles> Yes. >Apr 02 19:32:52 <D-Cypell> on the other hand... if I want the DES digest of a 100MB file I dont want to upload it to a website >Apr 02 19:32:58 <snarkles> True. >Apr 02 19:33:08 <snarkles> So we could draw the line there.. no file upload capabilities in the web version. >Apr 02 19:33:19 <snarkles> And no "key generation" capabilities. >Apr 02 19:33:30 <D-Cypell> That sounds like a good assertion to me >Apr 02 19:33:31 <snarkles> "If you want these fancy features, please download our desktop version." >Apr 02 19:33:39 <D-Cypell> Yes... exactly >Apr 02 19:33:44 <snarkles> This goes counter to what CyberTNT was proposing though. >Apr 02 19:33:50 <socrat> our "FREE" desktop version >Apr 02 19:33:51 <snarkles> Which was to only maintain a single version of the source. >Apr 02 19:33:54 <Cybertnt> ? >Apr 02 19:34:12 * Cybertnt doesnt care >Apr 02 19:34:20 <snarkles> And then when we do a build, automatically generate a desktop and web version from the same src. >Apr 02 19:34:27 <Cybertnt> i am like half away :/ >Apr 02 19:34:36 <snarkles> No, I know. But this is a discussion we've had before. :) >Apr 02 19:34:38 <D-Cypell> I think that is problematic... I like the idea and im a big reuse fan, but Java isnt going to work well on most standard web hosts and PHP isnt gonna work with for desktop apps >Apr 02 19:35:07 <snarkles> Ok, then that's fine. If Java is "not the tool" for simple web apps without a hell of a lot of work, then it's very easy to draw the line. >Apr 02 19:35:40 <snarkles> PHP 5 actually would correspond fairly well with Java. >Apr 02 19:35:47 <snarkles> But unfortunately PHP 5 isn't supported very well. :\ >Apr 02 19:35:59 <snarkles> So it'll be some work to get a PHP 4 version of what we have now in terms of class design, etc. >Apr 02 19:36:12 <snarkles> Cos PHP 4's OO capabilities are an absolute disaster. lol >Apr 02 19:36:38 <snarkles> The other option is Perl, whcih would be equally, if not more supported than PHP. >Apr 02 19:36:44 <snarkles> And comet keeps pushing for that. ;) >Apr 02 19:37:22 <snarkles> That involves an extra couple steps for the user (noting the path to the perl interpreter, chmoding the file so it's executable, etc.) but might be a way forward. >Apr 02 19:37:24 <socrat> no , perl is a dirty little thing >Apr 02 19:37:30 <snarkles> lol >Apr 02 19:37:35 <D-Cypell> Well it seems that PHP and Perl would compete for the web space. I lean towards PHP because thats what the existing script is >Apr 02 19:37:53 <snarkles> Yeah. It's also a language that's suited specifically for web, which is what we're gearing toward. >Apr 02 19:38:06 <snarkles> And the "upload and it's done" thing is a huge boon to usability. >Apr 02 19:38:25 <D-Cypell> I just had a thought.... if we wanna be really clever... >Apr 02 19:38:30 <D-Cypell> hehehe >Apr 02 19:38:45 <snarkles> Also, in that impromptu skills assessment, PHP was the only language other than Java that almost everyone had some skill in, which means it would be relatively easy to get people to help with it, etc. >Apr 02 19:38:46 <snarkles> Oh yeah? :) >Apr 02 19:38:50 <D-Cypell> The more trivial schemes can be implemented in PHP and placed on the web server >Apr 02 19:38:57 <snarkles> ! >Apr 02 19:38:59 <snarkles> Web services? :D >Apr 02 19:39:05 <D-Cypell> YES!!! >Apr 02 19:39:07 <D-Cypell> You beat me to it >Apr 02 19:39:11 <snarkles> LOL Sorry. :) >Apr 02 19:39:14 <D-Cypell> heh >Apr 02 19:39:23 <snarkles> That's kind of creepy. ;) I'll stop that mind-reading thing now. ;) >Apr 02 19:39:33 <D-Cypell> The client can consume a service for the straightforward stuff and you get the reuse >Apr 02 19:39:41 <D-Cypell> and the more complex can be native in the client >Apr 02 19:39:53 <snarkles> And the web service would be something we'd just house at a central place (sneak.cyberarmy.net or whatever) >Apr 02 19:40:06 <snarkles> we wouldn't have to distribute that to clients.. or would we? >Apr 02 19:40:11 <D-Cypell> Sure... can be very simple XML too. SOAP would probably be overkill >Apr 02 19:40:14 <snarkles> I don't know a LOT about web services, have only done some consuming. >Apr 02 19:40:22 <snarkles> Are you familiar with REST? >Apr 02 19:40:33 * D-Cypell shakes head >Apr 02 19:40:58 <snarkles> K, I'm not either. lol But it's what Amazon uses as one of their WS methods.. basically, you call a URL with the parameters you want to use and it returns XML based on that. >Apr 02 19:41:00 <snarkles> So like: >Apr 02 19:41:09 <snarkles> xml.amazon.com?isbn=12456557676 >Apr 02 19:41:23 <snarkles> Would give you back XML for the book containing title, author,e tc. >Apr 02 19:41:32 <snarkles> (that's simplifying it a bit, but that's the jist of how it works :P) >Apr 02 19:41:40 <D-Cypell> Thats perfect. Thats all that is really needed >Apr 02 19:42:07 <snarkles> Now the only thing about that option is it sort of eliminates the "offline" capabilities of our offline desktop version (for simple stuff). >Apr 02 19:42:23 <snarkles> Can I just ask at this point why we don't want to have the simple things in our Java version? >Apr 02 19:42:40 <D-Cypell> It would be possible to implement the standard web screens using the same functionality and use XSLT to get the 'prettys' >Apr 02 19:43:28 <D-Cypell> snarks: Thats true. We need to consider how big an issue that is. >Apr 02 19:44:01 <snarkles> Well, what is the issue exactly..? >Apr 02 19:44:03 <D-Cypell> snarks: Well, i dont know that we dont >Apr 02 19:44:21 <D-Cypell> snarks: sorry that was in answer to the 'no more 100% offline' >Apr 02 19:44:28 <snarkles> Basically that we're essentially treating all algorithms as "equal" under this design? >Apr 02 19:44:38 <snarkles> Yeah, got it. >Apr 02 19:46:04 <D-Cypell> Well... we certainly can implement all algorithms in both places. Its extra work, and it would probably be a good idea to set priorities for the first release >Apr 02 19:46:16 <D-Cypell> Maybe do 4 different strong crypto functions end to end >Apr 02 19:46:29 <D-Cypell> and then build out horizontally >Apr 02 19:46:35 <snarkles> OK, fair enough. >Apr 02 19:46:51 <snarkles> Pig latin, l33tsp34k, I could probably whip off right now in about 10 minutes though,. >Apr 02 19:47:00 <snarkles> So I don't see them as being a lot of extra work, >Apr 02 19:47:06 <snarkles> unless I'm missing something though. >Apr 02 19:47:38 <snarkles> The couple I'm not quite sure how to pull off is URLEncode and HTMLEntities, since that's taking a single character and converting it to multiple. >Apr 02 19:47:52 <snarkles> But I definitely don't want to lose those because I use them all the time for random things. >Apr 02 19:48:01 <snarkles> (encrypting e-mail addresses, etc.) >Apr 02 19:48:45 <snarkles> Or are you referring to the fact that the design is somewhat "generalized" so that it will encompass everything from Pig Latin to Blowfish? >Apr 02 19:49:35 <D-Cypell> I wonder whether it is a little over generalized, and over generalization can be a killer when you get towards the end of projects >Apr 02 19:50:11 <D-Cypell> It is better to do 2 or 3 things than to not do 50 :o) >Apr 02 19:50:23 <D-Cypell> but i dont know >Apr 02 19:50:38 <D-Cypell> Its hard to tell because I have not really dont this kind of thing before >Apr 02 19:50:54 <snarkles> Yeah I think we're all in the same boat there. ;) >Apr 02 19:50:58 <snarkles> Personally, though, I like the design. >Apr 02 19:51:03 <D-Cypell> I wrote a PGP client once, just for kicks, but ive not really attempted a large scale cryptography framework >Apr 02 19:51:09 <snarkles> Cos if there's one thing that changes constantly it's what crypt types SNEAK handles >Apr 02 19:51:24 <snarkles> It started out just being base64 and URLEncode, because I was doing like Try2Hack at the time. >Apr 02 19:51:33 <snarkles> Then I ran into another one which made me do something else, so I added that. >Apr 02 19:51:54 <snarkles> Then I ran into a situation where I needed to create a .htpasswd file and had no shell access to run htpasswd command. So I created that functionality. etc. >Apr 02 19:52:28 <snarkles> So by the design being generalized like that, it makes it so I can add anything that comes up very easily, by just implenenting a couple interfaces, and I'm done. >Apr 02 19:52:41 <D-Cypell> Agreed >Apr 02 19:52:52 <D-Cypell> Thats cool with me >Apr 02 19:52:56 <snarkles> As opposed to the version now, where I have to change 1. the drop down, 2. the function list, 3. the big case statement that decides which function to call, etc. >Apr 02 19:53:47 <snarkles> But I do kind of see what you're saying. >Apr 02 19:54:04 <snarkles> I thought the more "low-level" classes like Cipher and Hash were to help differntiate the different encryption types tho? >Apr 02 19:55:02 <snarkles> Or are you referring even more to the "big picture" scope of this thing eventually handling file/network encryption, etc.? >Apr 02 19:56:11 <D-Cypell> I dont really know. It so hard to rationalize this stuff because its just based on past experience and had become more a gut feeling than anything else. >Apr 02 19:56:23 <snarkles> Heh, yeah I hear ya. :) >Apr 02 19:56:31 <D-Cypell> The thing with generalization is that everything you 'allow for' has to be 'considered for' in the future. >Apr 02 19:56:39 <snarkles> But basically your concern is about generalization of the design, yes? >Apr 02 19:56:52 <snarkles> Hm. Could you explain? >Apr 02 19:57:00 <D-Cypell> We we design the gui.... we have to be always thinking "How will this work for....X Y or Z" >Apr 02 19:57:06 <D-Cypell> That is the cost >Apr 02 19:57:09 <snarkles> Ah, ok. >Apr 02 19:57:16 <snarkles> Yeah, that makes sense. >Apr 02 19:57:24 <D-Cypell> The flexibility of generalization doesnt ever come for free >Apr 02 19:57:38 <snarkles> This is how I pictured the development of this app. And you can tell me right now if I'm completely offbase and such (better to find out now ;)) >Apr 02 19:57:40 <D-Cypell> The XP school of thought is to do what you NEED today and worry about what might happen when it does >Apr 02 19:58:13 <D-Cypell> snarks: no... there is never a right or wrong answer to this. There are definate merits in the way it is currently designed >Apr 02 19:58:25 <snarkles> Version 1.0: More or less a carbon-copy of the current SNEAK script that does everything the current version does, only without the requirement for a web connection >Apr 02 19:58:36 <snarkles> Version 2.0: Implementation of higher-level and additional crypto tools >Apr 02 19:58:44 <snarkles> Version 3.0: Ability to handle files/networking/etc. >Apr 02 19:58:48 <snarkles> Version 4.0: Whatever. >Apr 02 19:58:54 <D-Cypell> Ok thats good because that sets a good roadmap >Apr 02 19:59:08 <snarkles> I think you're right re: your comment about XP. >Apr 02 19:59:13 <snarkles> That's the way I've been taught to do things as well. >Apr 02 19:59:47 <snarkles> I think that we just had the advantage of having somewhat of a "roadmap" and altered the design that way to start. >Apr 02 19:59:57 <snarkles> Which I don't see as a "bad" thing, although it's probably unconventional. >Apr 02 20:00:15 <D-Cypell> The problem with the XP approch though is it doesnt always mesh that well with some time honoured OOP concepts. Its natural to abstract and generalize in OOP. So the challenge for the OOP XP programmer is to find the line. And that will be different for everyone >Apr 02 20:00:26 <snarkles> Yeah, true. >Apr 02 20:00:45 <-- Cybertnt has quit (Ping timeout) >Apr 02 20:01:02 <snarkles> If this was your XP project at work, how would you guys have handled it? Probably just a simple "CryptMethod" class and all the various algorithms inheriting directly from that, or..? >Apr 02 20:01:35 <snarkles> And then as the need for other crypt tools (say, PGP key generation, network file stream encryption, etc.) came about, started abstracting the design, including more interfaces and stuff etc? >Apr 02 20:01:56 <D-Cypell> That is a difficult question to answer because usually what happens is exactly what we are doing now. Each person throws in their opinion and we wittle until we get a good answer. >Apr 02 20:02:08 <snarkles> Since I've only done team development in school, I'm honestly pretty ignorant when it comes to this kind of stuff. >Apr 02 20:02:22 <snarkles> Ah, well good to know we're not that far off the mark. ;) >Apr 02 20:02:30 <D-Cypell> Not at all :o) >Apr 02 20:02:42 <D-Cypell> Its just a little slower in text >Apr 02 20:02:47 <D-Cypell> and less black eyes ;o) >Apr 02 20:02:50 <snarkles> lol yeah. And without whiteboards, etc. ;) >Apr 02 20:02:51 <snarkles> LOL! >Apr 02 20:03:18 <snarkles> How about when this project is "done" we start on a collaborative development tool that lets you punch each other in the eye? ;) >Apr 02 20:03:40 <D-Cypell> heh a new set of MIME types >Apr 02 20:03:46 <D-Cypell> assault/punch >Apr 02 20:03:49 <snarkles> LOL!! >Apr 02 20:03:51 <D-Cypell> assault/kick >Apr 02 20:03:56 * D-Cypell chuckles >Apr 02 20:04:01 <snarkles> assault/britney_spears >Apr 02 20:04:08 <D-Cypell> LMAO!! >Apr 02 20:04:15 <snarkles> (reserved only for the most violent confrontations ;)) >Apr 02 20:04:55 <snarkles> Your other concern too was that it's more easy to abstract a design that's too specific than "specify" a design that's too abstract, yes? >Apr 02 20:06:02 <snarkles> brb one sec >Apr 02 20:06:08 <D-Cypell> Ok >Apr 02 20:06:53 <D-Cypell> But yes... absolutely. My IDE has many refactoring tools. One is 'extract interface/superclass'..... One is not 'Consolidate hierarchy' >Apr 02 20:08:51 <snarkles> Ha, indeed. :) >Apr 02 20:09:21 <D-Cypell> I think that replicating the functionality of the current sneak script for V1 is a great idea >Apr 02 20:09:29 <D-Cypell> that sets some nice, clear boundaries >Apr 02 20:09:56 <snarkles> Ok, cool. >Apr 02 20:10:11 <D-Cypell> I also think that the current design works quite well for that >Apr 02 20:10:25 <snarkles> Version 2.0 is still a little fuzzy for me, but I'd like to maybe take a round of peoples' suggestions and go from there. >Apr 02 20:10:34 <snarkles> And it's a moot point until we get 1.0 under our belts anyway. ;P >Apr 02 20:10:35 <D-Cypell> We can probably drop the AlgorithmProperty stuff too >Apr 02 20:10:53 <snarkles> k one sec, I have to look at this again cos I forget. lol >Apr 02 20:12:43 <snarkles> K, that's basically for sending in parameters such as "salt" to DES/MD5 functions, or..? >Apr 02 20:12:54 <snarkles> (I mean methods when I say functions ignore me and my procedural ways :P) >Apr 02 20:14:22 <D-Cypell> Yes, the idea was that it could be used to provide 'extra' configuration info for an algorithm. A salt would be one thing, a RSA key another >Apr 02 20:14:24 <D-Cypell> for example >Apr 02 20:14:58 <D-Cypell> I guess there is no need to remove it >Apr 02 20:15:11 <D-Cypell> but it probably wont be used too much in v1 >Apr 02 20:15:16 <snarkles> Ok. >Apr 02 20:15:24 <D-Cypell> at least the existing script doesnt need it >Apr 02 20:15:31 <snarkles> I'd prefer to leave it if possibl.e >Apr 02 20:15:33 <snarkles> possible* >Apr 02 20:15:42 <snarkles> Because otherwise we have to go and change the damn class diagram again. lol ;) >Apr 02 20:16:02 <D-Cypell> heh ok >Apr 02 20:16:12 <snarkles> (I know that's a stupid reason for keeping something, but if it's not hurting anything, then.. :)) >Apr 02 20:16:19 <D-Cypell> Ok thats great. So we have a nice clear scope >Apr 02 20:16:24 <snarkles> Ok cool. >Apr 02 20:16:34 <snarkles> Are you ok with that scope? >Apr 02 20:16:39 <snarkles> i.e. does it make sense, etc.? >Apr 02 20:16:45 <D-Cypell> I think its excellent >Apr 02 20:16:51 <snarkles> Oh, spiffy! lol >Apr 02 20:17:07 <snarkles> I'll make a post to the effect of "development roadmap" to get everyone on the same page. >Apr 02 20:17:33 <snarkles> And layout 1.0, 2.0, 3.0... >Apr 02 20:17:40 <snarkles> ANything else you think I should be covering? >Apr 02 20:18:33 <snarkles> I'll also throw the logs there for the folks who missed it (cybertnt, saj, etc.) >Apr 02 20:18:35 <D-Cypell> I dont think so. >Apr 02 20:18:45 <D-Cypell> Ok great
Signature:
Optional Image Link:
http://

CyberArmy::Forum v0.6
Generated In 0.01031 seconds