The one thing that hit me though is that this would possibly be a great platform for end users to use Cabernet. I’ve been concerned with getting the average wine drinker to install a database and Ruby on Rails and all that, but this is so elegant that I could imagine people buying pre-loaded units with Cabernet pre-installed- just plug in and go!

• Tcpreplay Pretty much in total maintenance mode right now. I’m more then happy to help users and fix any bugs which people report, but I’m not working on any new features right now.
• OpenPacket Looks like OpenPacket is in the process of being restarted and is going to move from Ruby on Rails to PHP. My last experience with PHP (about 7 years ago) was enough to make me swear it off and I’m just not interested in re-learning it again. Hence, I suspect my role will be limited to providing suggestions for the team.
• Cabernet Due to the move & my vacation to the Yukon/BC/Alaska, development has definitely slowed down, but now that things are getting back to normal I’m expecting to get some more work done. One thing I’m considering is starting a rewrite. Cabernet was my very first RoR application and I’ve definitely learned a lot and there are a lot of new and interesting plugins out there which are quite interesting.

  The other option is working on the UI. There are still a lot of things which need work (especially adding bottles and searching) which I don’t think would need a complete rewrite. I still need to decide what Cabernet should become- a OSS application for end users or a hosted solution (which seems to be the norm for the wine drinking community).

I first looked at PDF::Writer, but OMG- what a complex and confusing API. I just wanted something simple. Just my luck, a new PDF library is in the works called Prawn. Unlike PDF::Writer, Prawn is very simple to learn and use. I was able to generate a pretty respectable multi-page table in less then 30 minutes.

Anyways, the feature is a bit blue collar right now. Since generating PDF’s takes more then a few seconds, I really should have some ajax status thingy and use BackgroundRb for generation, but for a quick hack for our move Prawn made creating this feature quick and easy. Hence, if you’re looking to generate PDF’s in Ruby, I’d highly recommend you try out Prawn even though it’s still in early development.

Ultrasphinx is known for better scalability and reliability compared to Ferret, which is critical if Cabernet is going to go multi-user. It also should make it much easier to sort and paginate search results. However, things are a bit semi-broken right now as I need to completely rewrite both the views & controller to utilize the new sorting functionality of Ultrasphinx. I’m also planning on moving to a more RESTful design making the index and search the same view/controller method.

As a side note, I announced Cabernet on the Vino Cellar forums. I have to say I was a bit surprised with the complete lack of any useful feedback. Seems like everyone there is in love with CellarTracker and couldn’t be bothered to check out the demo before commenting and so ended up just asking stupid questions. *sigh*

The good news is that it sounds like Eric from CT is getting a lot of requests from people like me looking to interact with the CT site via a remote API. Currently Eric has implemented displaying some info via XML, but there’s no way to create or update records. Hopefully he’ll have something in a few months, but until then I’m going to keep hacking on Cabernet.

I have two reasons for this article: 1) to comment on some problems I’ve seen with some projects and 2) to remind myself of these things as I work on publicly launching the Cabernet project.

• Be nice!

  People will give you a lot of slack if you’re nice to them. When you’re trying to attract users and contributors, you should remember to treat others as you’d like to be treated. Just because you’re providing a (free?) service, doesn’t mean you can treat them poorly.

• Be responsive!

  If someone asks a question, reply quickly. Replying to questions and comments quickly creates a strong sense of having a conversation with the person and lets them know you see them as important.

• Talk is cheap!

  If you make a commitment to do something, follow through. Don’t say you’re going to do something and then don’t do it. It also means under promise and over deliver. I’m constantly amazed how happy people are when I set their expectations low and then deliver more.

• Communicate!

  Yes talk is cheap, but sometimes that’s all you’ve got. If for some reason something is taking longer then it should, you made a commitment you can’t follow through on or can’t reply in a timely manner, write a quick email/blog/whatever letting people know things are taking longer then you had hoped and when you think you can get back to them with the answer/solution/whatever. Nobody likes being kept in the dark.

• Accept their help!

  If someone wishes to help you out, then don’t look the gift horse in the mouth. Remember, most people are going to be leechers, so do everything you can to get them engaged and involved as quickly as possible. The more engaged and involved someone is, the more likely they’ll stay around and be contributing member of the community.

• Make sure you have time!

  I had actually originally forgotten to include this point, but it’s probably the most important. You need to have time. Running a website or any project for that matter requires more time and effort then you think it will- especially in the beginning as you try to drum up support and users. Dunno what to say here, other then look before you leap.

A big props to everyone at Dynamic Drive from where I’m happy to admit I stole CSS styles for the main and sub-menus. I’ve also made a number of tweaks to try to have a sensible color theme which should help with how the pages show on the eyes. Considering I’m well known for making the most ugliest web UI’s known to man (yes, I wrote the OneSecure/NetScreen IDP ACM), I consider this a big win.

I’ve also added a lot of code behind the scenes for user/group/permission administration- I’m still trying to figure out what exactly I’m going to do with the whole multi-user thing, but it’s good to allow users to be able to create accounts, login and be a part of groups and assign permissions.

Anyways, I still need to decide what nitch in the market Cabernet should try to fill. Being a RoR application, means it’s very portable, but also harder for most people to deploy. I could also go the more centralized hosting route, but I’m not sure I want to run a website for a bunch of strangers in my free time… not to mention the cost associated with doing so would be out of pocket. I guess I need to find some online wine forums and see what kind of interest there is.

That said, I’ve also taken the time to bang out a few new features and bug fixes. Tracking what wine you drink is now done via the Bottles screen (since you drink bottles not arbitrary wines) and when drinking or removing a bottle, you get a nice reminder where the bottle came from- great for people like me who have poor short term memory.

Also, I’ve been looking at Passenger, a simple to deploy Rails environment for Apache. Right now though, it seems that Rails applications are difficult to package up and deploy. Hopefully with Passenger and some simple scripting, I can come up with a nice solution and make a beta available soon.

Of course, the one big issue with Cabernet is that it’s about as attractive as Jabba the Hut. Not sure how I’m going to fix that, since it’s clear this is one area that I suck pretty badly at. Right now I’m hoping the beta gets enough interest that someone comes along to help. If you’re interested in showing off your CSS-fu, then let me know!

Yes, once again, I’ve proven to the world that nobody should ever ask me to develop a UI. (Last I heard, Juniper is still using the ACM for the IDP product. WTF?) I can’t really explain why, but clearly I have a bad eye for user interfaces and I fail horribly at basic implementation. I can’t even copy & paste correctly (obvious since the menus which I copied from Trac don’t look right). *sigh* Anyways, hopefully porting to 2.1 will give me the incentive to clean the code up a little, release it to the world and maybe get a few other people using it. If I’m lucky, I might even find a UI person to help out. (hint, hint, hint)

This is actually very relevant to me right now with Tcpreplay. As the project has become more mature and stable, I’ve started thinking about what to do next. But so far, I haven’t really thought about the “big picture”. Up until now, development has been very iterative: identify a new feature, develop, release, fix any bugs… wash, rinse, repeat. There really hasn’t been much vision, other then perhaps “Develop tools for using pcap captures to test inline security devices like firewalls and IPS’s”. Honestly, this was interesting to me back in 2002, but now that I’ve left the IPS vendor world it’s not that interesting and I’m pretty much out of touch/disenfranchised with the whole IPS thing.

The result is that I very rarely use Tcpreplay myself. On the other-hand, writing Tcpreplay has taught me a lot about software development and developing in C.

So what’s next then?

I’ve been trying to come up with what should be Tcpreplay’s “mission statement”, but honestly, I find myself not sufficiently motivated to see any of these grand ideas to completion. The result is I’m seriously thinking of putting Tcpreplay on the shelf for a few months (after 3.3.2) and either work on other projects like Cabernet which I actually use on a regular basis or start something new.