So you're telling me easy Jive->Open Data Format (and then into whatever FLOSS system you want) would be decent business?

Maybe some day people will realize that even with pretty features and cool looking stuff at the end of the day losing control over your data is a bad idea. This theme sounds very similar to my ERP days. Still tons of money to be made on the simple promise of "you'll always get your data and if you're not happy with us we'll make it easy to get it back/out".
I think the biggest problem that good FLOSS4business companies have is that they don't hire good enough sales people. Due to company culture sales is often seen as "the evil dudes" instead of "team awesome that can get our solution through the moats of corporate lolship and once it's there we can pretty much cruise because they'll be shocked at how awesome it actually is"

Given the situation it's a net+ that it's Java/.net. If he wants to stay in Ohio I think that's the languages with the best job options.

p.s.: I'd go with Java since the job market reach is bigger and there's cooler things for the JVM in general eventhough that can be argued (+you can expand into Android development which is the biggest plus). Java is still strong and doing very well in benchmarks as backend techology for web stuff, duno about the .net ecosystem (it's a bit too much lock in, they often assume SQL server etc.).
There's also more/better learning materials on all sorts of things for the Java ecosystem.
Ultimately the difference is pretty marginal, search around for companies you'd like to work for post camp and see what they are looking for.

Counterpoint being that my current estimation of skill/direction is that Microsoft>Oracle and I'll gladly admit that Microsoft has some of the best language talent out there. Thankfully Java is quite robust without Oracle.

I generally agree but a lot of people can't learn by themselves. The curse of knowledge also makes it pretty hard to imagine what it must be like to start from absolute 0 if you come in as a baker or something.
I think the university system isn't ideal because programming is essentially foundations+lots of practice. I'm still surprised companies haven't tried their own bootcamps, I belive we had this discussion in this thread before (Google does it now with Android).
There's going to be some middle ground eventually.

Teaching again and being lucky enough to have a 1st semester course with a lot of people with no programming background (somewhat surprising for university entrants but it's a strange degree that is a mixure of CS and design) has helped me quite a bit. I think I can create a decent "power/fastlearning" course for Java after this semester. I'm not interested in getting rich of it, I'll probably offer it for free but I think there's a decent market for a "minibootcamp" for people entering university in Germany (basically get you to a level where you can program and CS101 will be easy cruising).
I'm forced to come up with good metaphors for stuff that seems trivial for me. Using jars with labels and different formfactors for variables was extremly well received for example. For the next semester I'm going to use actual physical jars. Been thinking about stuff like that quite a bit, it's fun

The idealist in me is pretty bothered by the fact that German HS graduates are mediocre at Englisch and can't program. I think both are essential basic skills for the future and that's stuff I want to fix (so yeah if I ever wander into startup land it'll be to fix this stuff).

Java is good enough (Java8 is even decent). Like I said I think the difference isn't huge but here's why I'd go with Java:
1) The JVM is really good and optimized, good enough for you to not care (+not care if .net is slightly better/worse).
2) Java is more open. While .net is open sourced the entire Java ecosystem is more open.
3) Easier to find info/answers. Bigger "commuinity" (I could be wrong, has been a while since I worked with C#)
4) Following up on 2, web development with Java has a lot less lock in. This might change for .net in the near future but I'll always pick the option that defaults to PostgreSQL over SQLServer. Tomcat is pretty good, there's
5) Android (yay)

this is a huge it depends. i think you could cobble together something reasonable if you are basically piecing together pre built components in rails or something. if you are truly doing everything from scratch, i think even 3 months full time would be pushing it pretty hard. and ofc, there's a huge difference between a prototype and a production ready system. so your success would also depend largely on how loose your definition of "prototype" is.

with all this stuff, the real question is: does it excite you and would you enjoy it? if yes, just go for it, cause you'll get a lot out of it regardless of the degree of success.

Thanks! I got excited after reading stuff like this; although this guy had the HTML/CSS background (but OTOH I have a better tech/programming background).

I def think I would enjoy the process, and if I ever want to get into the tech startup world (I do), being able to build stuff myself seems like a huge asset. And for this specific project, it's either gonna be something DIY or nothing at all, as we don't really have the means to get outside developers/designers on board.

I guess the main constraint is gonna be 'time', as the year is already pretty busy as it is.

Short question to make sure I'm not missing something obvious. There's a Python library I use with a method

The naming of the keyword argument strange_name makes no sense and there's an obvious one that would be better. As far as I remember you can't simply do
Is there any Pythonic way to get this done or would I have to lobby the authors to change their API and break old code?

Basically...change an existing method signature without breaking backwards compatibility (something you could hack up with overloading in other languages). I feel like I've actually done this before but can't think of what to search for.

This is definitely a big headache for the mud rakers, but I don't think the higher ups know about it or really care. They just know things work, and don't know that the whole entire company's critical data is stored in piecemeal Excel sheets with 100M LOC of unreadable VBA scripts, half of which are corrupted.

On the other hand, many OOS companies are making pretty good money with ETL.

Plus there is a load of confusing and plain wrong information out there. For example, the Stanford Database class was brought up earlier. I no longer suggest that class, but I don't know a good alternative resource.

The codecademy sql class gives you pretty much as much info I think and is way quicker and easier to absorb. May not cover normalization, but to most users of sql that's probably not a big concern (compared to db designers)

I think this would work:

Yeah that should work and looks like what I did before, thx. My brain could only remember "there was some renaming involved"

What would I do if I were the author of some library that exposes awesome_method and it is in use but now I want to rename one keyword argument without breaking backwards compability (just wrap somehow using that suggestion)?
Seems like a use case that can come up some times? I guess you can always just change the keyword argument and communicate that but maybe sometimes you don't want to do it?
Genuinely curious how this is handled in Python or if I'm inventing problems when there are none
[monkey patch?]

I imagine you could just add the keyword argument (with a default value it's optional) no?

I disagree with this viewpoint.

How is said self-learner going to learn to select from a non-existent database? He has to create his own database first, right?

No, he can run a script that creates the database for him. There are certainly career paths that will lead to never creating a database, probably more than lead to creating them.

Normalization is pretty important to cover imo, if only because it would reduce the amount of times I have to deal with horrendous non-normalized databases. Frankly I'd settle for just never seeing a 1NF violation in the wild ever again.

accidentally overwritten windows =(

lesson learnt:
always backup
always double check which installer option u selected

rip data, rip $119 for a new copy of windows

also is it worth trying to pirate a copy of windows?, or am i just asking for a world of pain?

Am I misremembering, or are we talking about the guy who wants to do a mockup for an MBA class?

Or was there a separate guy interested in self-learning?

Why do you think you need to buy a new copy of Windows or pirate it? Just use the same key, it will work just fine.

Might I recommend clonezilla as a fantastic and easy to use imaging backup solution?

Gets a little funkier on machines with newer preboot environment software (I think UEFI) due to keying requirements to boot from OS.

Also could make the switch to Linux now and never look back :- )

Why bother learning SQL at all if you can just ORM your way to glory?

Plus normalization is easier to learn than select queries, IMO.

I did once use a byte with mapped bits to represent which day of the weeks a thing was active for. Am I bad and should feel bad?

And on the eighth day, he rested, and gazed upon his extra bit, and knew that it was good.

Am I misunderstanding, a field that allows null values means the table is not normalized?

From the wiki it sounds like "maybe"...

Yeah I checked out wiki and another article and couldn't get a clear answer.

Guess it's "contested"

I don't know, seems a little cumbersome to have a table for every piece of user data that could be null.