[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: obsidianmusic alternative? (music library with streaming or download from webserver)



That would depend on the usage, consider this:
- find an old jukebox
- strip the old (defective?) hardware
- insert cheap modern computer hardware
- include touchscreen for song selection

I'm seriously considering something like that. The web interface would be ideal for parties were people can compose their own playlists, which
are then mixed with other peoples playlists. This way people can add
songs using a laptop or smartphone. I still need to work out some
details, but it seems like a good idea to me. In this case mpd would
shine.

100% true. And I must admit, such a jukebox could be a nice item to have :) And, nowadays, you could simply replace the screen by a graphic tablet, or find a cheaper touchscreen mounted on something like a raspberry pi. For the computer stuff, it could become something like 200-300€: 20 for the rPI, 80 for a hard disk, other for screen :)

Personally I think it's partly because of the poor quality of code some people write. I've seen some hopelessly ineffective code that could be
both easier to understand, shorter and execute faster (a lot).
Like replacing a all instances of a character with another one in a Java
String, using recursion (1 char replaced each call). It was dead slow
and gave out of memory errors (Java strings are immutable, so you keep
the whole string several times during the whole recursion thing).
A simple String.replaceAll() was all that was needed.

Ahaha nice one :D
Why doing something easy when you can obfuscate it...

If you're talking about servers, I couldn't agree more. But for
desktops, I'm a bit more relaxed. Still I prefer good software that does
one thing only, and does it good/correct over software that does 101
things badly, so long as both work of course. No matter how good the
software is, if it doesn't fit the use case, it's useless.

Hum... it depends on the computer. I'm quite short on money, so I take care to buy cheap hardware. Cheap hardware often have lower performances, and here comes good softwares which does one thing. I could use some heavier tools, but my computers would become really slower (but I'm using some gnome tools, like meld and dia, because I did not found equivalent without gnome deps...) By example, I've bought a 1015 pem two years ago, immediately replacing windows by a debian. That computer, with low perfs, reply to me faster than most computers I can see, but someone which does not know how to use tiling wm and terminal emulator to start programs (and so, who does not know how to know which tools are installed) aka: normal user, would be unable to use it. On my side, I'm now faster with it that on most other computers I can use, and not only because of software's performances.

Another point good softwares doing only one thing if to learn tinkering, they are easier to combine with other tools, when their objective is to do *one* thing, because doing one thing is useless when you can not interact with the world correctly. But I use some IDEs, so I guess I can not say I'm a real UNIX extremist, because those tools are doing text edition, debugging, project file management... between others.

What about oggconvert? It's a gui to convert various media files to ogg.
It certainly works with wma files and relies (I think) on gstreamer.

Sounds like it can not : "video/x-ms-asf" is not supported. I'll investigate, now I have a starting point :)

Possibly, I always found it to be more lightweight on windows than
installing some random codec pack from a creepy website. I just use it on Linux because of the history I have with it on windows, never let me down. It is certainly possible to run it without a GUI, see the package
vlc-nox.

Yes, I know it is possible, even on windows, in facts. On my side, I've noticed most free tools I was using on windows have disappeared from my uses, for a reason or another. Miranda IM is not native (and using wine is not really something which makes me happy), notepad++ also (and there are other good enough text editors), vlc is nicely replaced, firefox has been thrown away... Still keeping code::blocks (C++ IDE)... but I'm currently writing tools to replace it (unsatisfied with some aspects, and I've discovered a new way to use my screen space this year... tiling window managers are now my way, and I think there is a lack of tools, so I'll write some). But all of those tools made the transition easier, each one with their own way: miranda learned me that to be complete to configure does not mean complex, notepad++ said me that there are nice text editors, other were just portable and lighter than their windows usual concurrents. Transition is easier when you keep tools you were using :)

I use a bit of both, I run testing so I'm not that far behind on
versions, a rare package is installed from testing, some from third
party sites, but not much.

I can name tools I have from external repos... opera (does not exists in non-free), code::blocks (nightlies are far more recent, and stable) and i3 (my window manager), to find bugs. When I'll have enough will, I'll try to contribute to that last one... No need to have other more recent. Globally, I run testing, without any problem, and recent enough softwares.
So, I agree with you :)

To give Java some credit, performance is actually quite good if the
developer of the application knows what he's doing. Due to JIT (just in time) compiling within the JVM you run code like you would with compiled C-code. And there are other optimizations, some java code will actually
run faster than the equivalent in C or C++, but to be honest that
depends entirely on the compiler.

Hum... I agree for some parts, as I've said on various places that the problem is more frequently the programmer than the language. About speed, I am curious to see a java binary running faster than a compiled (C or C++, by examples) equivalent. I've read some benchmarks claiming that... and C/C++ code were sometimes outdated, did not used basic optimizations, or stuff like that. Of course, a Java dev will not know C++ tricks, and a C++ dev will not know Java's ones. But, I can see no reason to say Java softwares can be faster than compiled one, since Java's ones needs to be, at least, recompiled one time each time the computer has been started, or each time the computer needs memory (quite often on my netbook... at each compilation ^^), and generally speaking will use an intermediate software. I would change my opinion if it would be possible to build a usual elf (linux exe) or exe file from a java one, because it would remove a layer, and the compiler would be able to use all optimizations of the particular processor, which can be made in C++, but I do not think people really take care about that (except gentoo users, of course :) ) But my main reason to like C++ is really the RAII thing, against garbage collectors in all fashion languages. It allows to clean resources when you really do not need them. It can became the most interesting when, by example, you are managing limited resources: you've spoken about ports previously. RAII is nice for network/databases/file stuffs, by example. The programmer never needs to say MyDB.free(), it is made automatically when the processor goes out of the scope MyDB was initialized.

Other reasons are operator overloading (use with care, of course), powerful template system, multiple inheritance (use with *real* care), possible usage of non OOP stuff (because, seriously, having main() as a class member, this is just stupid imho. And it's an example.)

But java have its strengths, too, like a richer standard library (were C++ one is very small, because designed to be used by almost each users, and so lacks features which would serve most users), no re-compilation needed between OSes (but, is it really useful? Compile time for releases is not really costly) and highly verbose language (useful to learn programming, I think... but I do not like it, since I'm more used to short lines) and, java is far easier to parse, giving more powerful tools (at least, I've found good analysis tools for C++ supporting the entire language), which improve writing speed (code-completion and automated static analysis are some examples...). Simply, those strengths are not enough for me (and for tools, to have used eclipse for something like 3 months has been enough... crashes when you are writing faster than the autocompletion is able to think is not acceptable. Netbeans was a bit better, and luckily I was able to use it in secret, or I would be currently mad!)

Already have, see jukebox ;)
I admit your idea of the jukebox have it's interests. I would imagine it in a pub with 1980's decoration theme... or how to mix modern features with styled stuff :)


Reply to: