An open-source alternative, Beagle provides fast indexing and searching of all your files. Kevin of the LinuxForums.org content development team has used it, abused it and reviewed it.
An open-source alternative, Beagle provides fast indexing and searching of all your files. Kevin of the LinuxForums.org content development team has used it, abused it and reviewed it.
I’ve been intrested in installing beagle, but heard the stock Fedora 4 kernel didn’t have inotify on. If that is the case, does anyone know if I’ll have problems with my Nvidia drivers if I rebuild my kernel? Would love to get Beagle on my system.
BTW, First time first post. 🙂
I have been using it with Ubuntu Breezy and it works quite well. Haven’t really had problems with stability. The only problem I hear from people is the huge memory usage. I haven’t experienced it myself with 0.12.
They have now added an indexing & search option under system prefrences of gnome which allows the user to set privacy settings (exclude directories from being indexed) or add other directories to the index list. By default only the home directory is indexed.
Yes, but he wants inotify. Which AFAICT isn’t enabled on Ubuntu either.
I think you’d be fine rebuilding with the inotify patch. But just build it as a second kernel boot option and then if it makes your nvidia drivers broken (which you may have to reinstall
) then you can just boot back to your old kernel. I don’t see why inotify would mess with anything that’d break video drivers though.
Also, you can use beagle without inotify. It’s just not nearly as cool.
… was invented long ago. It is a wonderful thing called ‘directories’.
Yes, and directories are such a great help when searching through the contents of files…
Ever heard of grep ?
Ever tried to grep through dozens of pdf files? How about .doc files? Image meta data?
> Ever tried to grep through dozens of pdf files? How about .doc files? Image meta data?
Good point, you are totally right 🙂 I never thought of that, but searching something in multiple PDF documents is obviously a good thing.
However, I still believe that throwing all of one’s files into a giant directory is quite a bad idea in the long term (I have seen such a behavior with many people, especially P2P users).
Yeah I have the bad habbit of leaving things in the same folder for so long… it gets all messy and ugly!
Beagle could incorporate a sorting function that would let you sort your files into folders as you find them and select them for sorting… it would be cool.
I agree, I also at least try to keep my stuff organized.
However, the great thing is, you can still keep your stuff organized the way you are used to and use beagle nontheless. 😉
find + pdf2latex + grep.
I think with those three, and a good knowledge of find and pdf2latex you’d be able to do it in one line. Something like:
find . -exec “pdf2latex | grep [keyword]” -name “*”
Or something along those lines.
Of course, installing all the latex stuff is a bit of a disk space commitment…
seth started this http://www.gnome.org/~seth/storage/ but i never saw any progress with it. i followed it from the time he annoucen..
seth where r u ?
Hmm. C#? Is this the beginning of the push for Gnome to use Mono instead of Java? Should I start learning C# now?
Hopefully. Mono seems to work well with Gnome, and quite a few useful applications are now written in that language. You’re free to use whatever [supported] language you’d like, of course.
There is nothing wrong with c# as far as legalities go though for political reasons it wont be included in Gnome core. This is actually a good thing because c# does have a real drawback in its huge memory usage and bloat (and the same also applies to java). Most c# desktop apps need in excess of 100MB ram to run so it would make gnome’s memory footprint rocket and currently its not really possible to run beagle or any other mono app on a 256mb system with gnome without either being in swap or perilously close to it.
IMO, Gnome needs to create its own high level language like a static gcc4 compiled ruby/python which is both fast, highly productive and non-bloating. There really is no excuse for using over bloated languages on the desktop as they can harm adoption of linux on older hardware and in particular developing nations.
objective-c looks like a good canditate to me. it is simple , dynamic like smalltalk (or python) and compiled with gcc. But maybe it needs a real garbage colection
hey i registered
IMO, Gnome needs to create its own high level language like a static gcc4 compiled ruby/python which is both fast, highly productive and non-bloating. There really is no excuse for using over bloated languages on the desktop as they can harm adoption of linux on older hardware and in particular developing nations.
i think objective-c meets most of your requirements
Objective-C isn’t an especially high-level language while managing to incur various performance overhead from its usual implementation of dynamic dispatch. And to be honest, all the creation of another language will suffice for is to reduce the number of people interested in developing for the platform; no one is going to want to learn a comparatively obscure language in order to work on GNOME.
I think rather than offer arm-chair development advice, people should continue to develop their software as they see fit. In the case of GNOME this is largely C with a touch of C++, Python, Ruby, C#, and the like. Unless people intend to actually make anything happen by actually doing things, this endless philosophizing about other people’s work is fruitless.
yes maybe but
-Objective-c already exists for gnome (like python and ruby)
-It is gcc compiled like like the poster i replied to wanted
-It’s runtime does not use to much memory like the poster i replied to wanted
-python and ruby aren’t especially high-level either
-i like to offer “arm-chair development advice” i am here for fun not to work!
-python and ruby aren’t especially high-level either
Where did you get that obviously uninformed piece of information from? Now, I love Objective-C/C as much as the next guy, but they just don’t compare to python and (from what I’ve seen) ruby.
Specifically, python shows the same elegance and power as C, but with more features, unlike C++/java/whatever.
-bytecoder
“-python and ruby aren’t especially high-level either
Where did you get that obviously uninformed piece of information from? Now, I love Objective-C/C as much as the next guy, but they just don’t compare to python and (from what I’ve seen) ruby.
Specifically, python shows the same elegance and power as C, but with more features, unlike C++/java/whatever. ”
Ok what do you believe high-level means? more features? elegance?
I think it has to do with a level of abstraction so to me procedural languages are about the same level,object orientated languages somewhat higher . SQL or Haskell are even higher level. I do not have a mathematical exact definition what high level means. I think it is debatable but i am probably not more uninformed as you are.
yeah, the author of the article was a little mistaken: Beagle won’t be included in core GNOME, because it’s a Mono application, and like Jamie said, Mono won’t be a core part of GNOME.
In this case the GNOME (and KDE) will dead. The .NET architecture with CLR, C#, dotnet class libs, visual studion, etc is a very-very strong weapon of the windows desktop and I can’t see any similar thing on the linux side. IMHO the windows will beat linux with the better developer environment.
If you want to. I don’t think java to c# will take you all that much time… Of course, you’ll be frantically looking things up when you first get to using it; but that’s the best way to learn (at least for me).
Learn Python, it’s more interesting (different) anyway
.
Or learn lisp or something.
Some people would scoff at the idea of building medium-to-large applications in a scripting language, but I don’t think that’s a bad idea, especially for the free software community. Python is probably the best bet right now. Check this out, it’s interesting:
http://www.mjtsai.com/blog/2002/11/25/perl_vs_python_vs_ruby
However, perl6 looks interesting. It has the ability to compile both to bytecode for Parrot or to an executable for a number of supported architectures. Perl tends on the line-noise side, though. It takes a concerted effort to write readable perl, and I hope perl6 is a step in the right direction in that regard.
Scripting languages are becoming more and more powerful, and there’s no reason why they can’t take on more extensive applications (pun intended).
1. Because without strong type system reduce the reilablity of your code. You can’t see it until your applications is not too big… The “dynamic typing” is not a too new thing, but the really big, enterprise-level applications always written in C++, Java, etc and not Perl, TCL, etc.
2. The dynamic languages always will slow and eats many memory. Python is one of the slowest thing under linux.
3. The language is not enought. You need a big, coherent, stable class library. Python has own class library, but it is not comparable to .NET.
1. Because the lack of strong type system… Sorry…
“1. Because without strong type system reduce the reilablity of your code. You can’t see it until your applications is not too big… The “dynamic typing” is not a too new thing, but the really big, enterprise-level applications always written in C++, Java, etc and not Perl, TCL, etc.”
The modern scripting languages support strong typing and scoping. Ruby and Perl are probably better than Python in this regard.
“2. The dynamic languages always will slow and eats many memory. Python is one of the slowest thing under linux.”
So C# and Java, falling into the dynamic languages category, will always be slow and memory hungry as well. I guess we’ll always be stuck with C/C++…
“3. The language is not enought. You need a big, coherent, stable class library. Python has own class library, but it is not comparable to .NET.”
This is why Python is the scripting language of choice right now. It has a remarkably large array of available libraries and bindings. It might not be as extensive as .NET or Java, but perhaps the nature of scripting languages makes library support a little less important?
Examples of primary candidates for scripting languages in free software: installers, package managers, configuration utilities, daemons (like beagle), etc.
“The modern scripting languages support strong typing and scoping.”
It is very nice if you create one-man project. But if anybody can use dynamic typing it will reduce your system stability.
“So C# and Java, falling into the dynamic languages category, will always be slow and memory hungry as well.”
Java and C# far faster then Python.
“It has a remarkably large array of available libraries and bindings.”
Coherent, stable class lib <> many separated small libraries and bindings.
“but perhaps the nature of scripting languages makes library support a little less important?”
Only if you want re-invent the weel in every new application 🙂
“Examples of primary candidates for scripting languages in free software: installers, package managers, configuration utilities”
Yes, and it is the main target of the scripting languages: small sysadmin scripts.
Yeh Gnome-Storage was a brillant idea, has it a future?
I here it should integrate in gnome-vfs but i’m not sure.
Will Gnome-Storage come up, beagle becomes superfluous. And that is not in Novells interest, they will push C#/Mono and beagle is the best and futurefull software they had in C#/Mono.
Poor review, but the demo videos he pointed at the end are nice.
As a spotlight user I confess it takes a while to stop feeling guilty for the mess in the HD though…
GDS 2.0 is truly integrated with windows
GDS 2.0 is 300k only (web server embedded sidebar indexer searcher )
GDS 2.0 have unique ranking system
GDS 2.0 plugins can be developed in python java c/c++ c#
GDS 2.0 is really useful
beagle is written in c# for sponsoring mono
beagle is slow doggy
beagle is not well integrated with gnome
beagle plugins must be developed in c#
beagle indexer is obscene
beagle interface for queries is obscene
GDS 2.0
beagle 0.12
…
Version numbers aren’t especially useful for comparison.
beagle is written in c# for sponsoring mono
agreed
beagle is slow doggy
you should try inotify version.
beagle is not well integrated with gnome
A simple modification of gtk would suffice. But, I hope it wouldn’t
beagle plugins must be developed in c#
ever heard of CILC?
beagle indexer is obscene
Lucene is not obscene
beagle interface for queries is obscene
Wouldn’t argue here
More or less problem here is memory consumption with sloppy memory handling in mono. After a day of two and almost any active .net application completely consumes your memory
I’m sure GDS was truly awesome when it was in early development too….
Maybe, but i’m sure they had many more people working on it and full time 7 hours a day.
People seem to forget opensource apps dont get nowhere near as much time and manpower devoted to them as non opensource apps.
I was being ironic
I’ve never understood the interest in Mono. Why encourage dependence on Microsoft’s technologies in the Open Source world? Most linux users I know are more interested in getting away from Microsoft than getting closer. If you want to run .Net and program in C#, why not just run Windows? It just seems like a really bad idea to support Mono and any apps written with it, like rolling the wooden horse inside the gates of Troy.
Because the C# and .NET is the far best language and environment (class libs, runtime, etc) to write desktop and server applications.
For server apps, ruby on rails trounces both J2EE and C#/ASP productivity wise.
For desktop apps, Delphi (or the open source variant Free Pascal) is a far better tool as it offers the same component based RAD but unlike c# and java is non-bloating, has higher performance and thanks to its smart linker has no external dependencies/frameworks.
Ruby on Rails isn’t in the same league as J2EE or ASP.NET. Comparing them as such is not sensible.
Further, Lazarus is not comparable either. Delphi is on the road to becoming a .NET product. Heh, in both cases you compare a much less complicated, much less finished, functionally incomparable technology to something else and claim them as superior. They are what they are.
Ruby on Rails isn’t in the same league as J2EE or ASP.NET. Comparing them as such is not sensible.
Well according to someone at IBM it is:
http://www-128.ibm.com/developerworks/linux/library/wa-rubyonrails/…
Further, Lazarus is not comparable either. Delphi is on the road to becoming a .NET product. Heh, in both cases you compare a much less complicated, much less finished, functionally incomparable technology to something else and claim them as superior. They are what they are.
Well as a professional I use the ordinary version of Delphi for all my windows desktop programming and it wipes the floor with both java and C#. Delphi is 10 years old mature technology and its a professional RAD tool unlike the amateurish immature C#. In fact the over hyped C# is only good for ASP stuff but on the desktop its a joke with unbelievable amounts of memory being consumed as well as painful start up times – my customers would never forgive me if I gave them that. With Delphi based apps, all my customers are very happy.
>.my customers would never forgive me if I gave them
>that. With Delphi based apps, all my customers are very
>happy.
Yeah, but you have to program in object pascal! I’d rather be investigating a persistant boil on Leroy Jenkens’ backside than going back to Pascal after learning real languages…
> Well according to someone at IBM it is:
Do you really think that
Aaron Rustad ([email protected]), Technical Architect, Anassina, Inc.
constitutes “someone at IBM?”
It’s almost like http://www.osnews.com/comment.php?news_id=11260 all over again.
> Well as a professional I use the ordinary version of
> Delphi for all my windows desktop programming and it
> wipes the floor with both java and C#.
The native Delphi doesn’t offer the same things that Java and .NET offer. I don’t really care what you personally like; Delphi is not a replacement.
> my customers would never forgive me if I gave them
> that.
Then don’t use it.
> With Delphi based apps, all my customers are very
> happy.
Not if you’re using Lazarus they’re not. I’m sure all of your MacOS and Linux customers love your programs. I bet you don’t actually have any of those, right?
The native Delphi doesn’t offer the same things that Java and .NET offer. I don’t really care what you personally like; Delphi is not a replacement.
Of course it is – you obviously have never used Delphi if you think that. Dont forget C# is basically delphi with c syntax and although the .Net framework is richer than Delphi’s VCL, Delphi has huge open source component libraries like http://www.delphi-jedi.org/ which make it just as rich if not richer.
I can write almost any app in delphi just as fast and easily as C#/java and it will be faster, use a fraction of the memory, have super fast start up time, be less bloated than an equivalent C++ app and need no distributables. So dont tell me C#/Java are superior on the desktop – they are a million miles away from Delphi in those areas.
“Of course it is – you obviously have never used Delphi if you think that.”
I am a professional Delphi (Kylix) developer, I wrote more then 100 000 lines in Object Pascal. And I know (a little bit) the .NET. Believe me, the two thing is totally different. Not the language, but the class library and the infrastructure.
And IMHO the Delphi is totally dead. From the D8 the Borland is focused on .NET, but IMHO the C#/Visual Studio 2k5 is better idea then Delphi for .NET.
Not the language, but the class library and the infrastructure.
Yes but Delphi has all the stuff you need for *desktop* apps. Window.system.forms is very similiar to the VCL. The only real difference is that .Net is more extensive in size and features but so is the VCL with third party components. I admit MS monopoly power and marketing may well spell doom for Delphi but it wont be because of technical inferiority.
And IMHO the Delphi is totally dead. From the D8 the Borland is focused on .NET, but IMHO the C#/Visual Studio 2k5 is better idea then Delphi for .NET.
Borland is committed to keeping a non .net version of the Delphi compiler for future versions (I use delphi 5 a lot so I dont bother with the later releases anyhow).
As you are probably aware, the financial handouts borland received from MS was conditional on them supporting .Net in Delphi but borland aint giving up on the native win32 compiler anytime soon.
Borland’s poor financial health is a worry and native Delphi may end up being neglected like Kylix but hopefully they will open source Delphi/Kylix before they snuff it.
“Yes but Delphi has all the stuff you need for *desktop* apps.”.
It is true. But it is often not enought. For example, if you want to create multi-tier application with ritch client and web client based on common business layer the Delphi is not the best platform.
“but so is the VCL with third party components”
The main problem is the stability. VCL a little bit buggy. The external components can be very buggy.
“Borland is committed to keeping a non .net version of the Delphi compiler for future versions”
Borland is created Kylix, CBuilder-X, etc. IMHO the commitments of Borland are not too reilable…
“but borland aint giving up on the native win32 compiler anytime soon”
IMHO if the .NET environment will be installed on the most of windows machine the ponderousness of win32 code will be minimal in the most of windows applications.
Thank you for demonstrating to me that you don’t understand anything that isn’t Delphi. Delphi does not provide the same functionality that .NET and Java do. It has absolutely nothing to do with whether they suit your needs for the applications that you develop more. I don’t care. They are not functionally identical. They do not accomplish the same goal. You ramble on about them being “better” simply because they do what you want. That’s great; Visual BASIC is the greatest programming language in the universe, even better than SQL, the other greatest programming language in the universe! They totally rock everything! GNOME should use VB! Yes, that is heaping load of sarcasm.
Further, your Delphi code is completely useless outside of Windows, because Lazarus is not in anything resembling a commercially viable state. Mono and Java will continue to develop on multiple platforms while a platform around Free Pascal continues to exist on the fringe unused in the wider development, where any discussion of GNOME is relevant.
I don’t particularly like C#, Java, or Delphi to be perfectly honest, but I at least have the honesty to assess their respective capabilities rather than talking out of my ass about how much one of them “owns” the others. Delphi is a million miles away from mattering to GNOME.
Duh!
I was merely pointing out that for *desktop* apps C# is not a better solution than Delphi (on windows). I dont doubt for example that C# and ASP is a nicer platform than Delphi for web apps (I know cause I have used it there) but on the desktop its a different matter. Why? Because its the quality of the app that counts and thats where Delphi owns Java and C# (the poor quality of C# and java apps on the desktop makes them unprofessional tools in that department IMO). If you look at my previous arguements in this particular thread you will have realised that.
I am asserting that theres no reason for a component based RAD langauge to be over bloated and Delphi is proof of that.
ps I never mentioned Gnome wrt Delphi or VB at all – where on earth did you get that from? Or are you just trolling?
> I was merely pointing out that for *desktop* apps C#
> is not a better solution than Delphi (on windows).
You’re claiming that Delphi is a better solution, and now you’re specifying a specific platform (one that is completely off-topic I might add). C# may or may not be a “better solution” than Delphi on Windows for a myriad of reasons.
> I am asserting that theres no reason for a component
> based RAD langauge to be over bloated and Delphi is
> proof of that.
The native Delphi (as opposed to the .NET incarnation that Borland is moving to) does not accomplish what Java or .NET do, which is why the nature of your assertions aren’t useful. That if you don’t use a virtual machine, real garbage collection, and assorted other things that you then don’t have to pay for them hardly comes as a surprise to anyone. If you said that instead of “Delphi owns Java and C#,” then I might say “duh” but I’d hardly object to it.
> ps I never mentioned Gnome wrt Delphi or VB at all –
> where on earth did you get that from?
When you look at the title of this article, what do you see?
Beagle Desktop Search Reviewed
What’s the subject of this discussion?
Why write a Linux app in a Microsoft language?
The point of mentioning VB was to pick another, much more popular language that meets the requirements of hordes of programmers for constructing their wares, and suggesting that because it meets their requirements that it is “better,” despite it not being equivalent to any of the three platforms (though being much closer to native Delphi development than say .NET or Java), and the putting it to use precisely in the environment being discussed (“Linux apps”) where it’s entirely useless.
I would typically rather use a Delphi program on Windows than one written in C# or Java, for much the same reasons that you’ve outlined. That doesn’t mean that Delphi is an example of .NET or Java done better, because it doesn’t accomplish the same things.
>For server apps, ruby on rails trounces both J2EE and C#/ASP productivity wise.
Yeah the big enterprises will replaces their old .NET and J2EE solutions to the new fantastic ruby on rail :-))))) If you see the job search sites, every company search ruby, python or TCL programmers…
>For desktop apps, Delphi (or the open source variant Free Pascal) is a far better tool as it offers the same component based RAD.
Delphi 2k5 is also .NET based, it can create win32 applications but this tool is focused to .NET. The FreePascal (with lazarus) only a very nice toy, but not useable for real application development (if your app is bigger then Hello World). And the VCL is also a very primitive toy if you compared to .NET or Java class libs.
The only one comparable technolgy to .NET the J2EE, but IMHO the .NET is better, and the future of .NET is more certain.
This “review” was done by a Beagle developer, not a neutral user. See the screenshot right in the middle, http://images.linuxforums.org/content/beagle.png
beagle is OS, you can get the sources, edit them, view them, compile them….so, where is the problem?
I can take the same shot without being a beagle developer
>>”The initial setup the the one that all of the worlds anti-linux zealots love, citing long dependency list, kernel compiles and modification of all sorts of system configurations.”
You aren’t a zealot if you don’t want to compile and install alpha code like Beagle. You aren’t a zealot if you don’t kow how to compile and install Beagle and don’t want to larn. You aren’t a zealot if you chose to avoid building your own software.
Identifying such people as “anti-linux zealots” is simply cult behavior intended to boost the egos of the alleged cognoscenti.
Beagle is what can happen when you don’t have a boss who set a deadline date: developers noodle and play with their code forever, never releasing a finished product.
Beagle shows real promise, I like it and a good review. Thanks Google for leaving Linux users in the dark, use Linux technology but dont give anything back, thanks, thanks alot. I guess Google dont want any more people to use there search especially since there competing with Microsoft, never mind your loss which you failed to capitalise on.
It’s amazing the brain dead out there who STILL think Linux is not worth the effort.
Beagle shows real promise, I like it and a good review. Thanks Google for leaving Linux users in the dark, use Linux technology but dont give anything back, thanks, thanks alot. I guess Google dont want any more people to use there search especially since there competing with Microsoft, never mind your loss which you failed to capitalise on.
It’s amazing the brain dead out there who STILL think Linux is not worth the effort.
Google’s whole infrastructure and search mechanism, that is to say the very core of their business operation, is built on linux. So are the google appliances they sell to enterprises for indexing their own web/intranet sites.
Given the portability and scalability of linux, how difficult would it be to take a theoretical Google Desktop Search for Linux and scale it to search intranet or even internet sites. See the conundrum?
They could also release it as a closed binary and restricted (free as in beer but not free as in speech) license to try and prevent people from deploying it in a competitive manner, but then we all know how the linux crowd feels about that, right?
So how does google win in a situation like that?
I’d love to see google desktop for linux. I won’t install Beagle (I’m kde and don’t want to have to install 100 some odd dependencies to install it) and Kat just confuses the hell out of me. But from a business point of view, I can understand why there may be a reluctance to release a desktop product for linux.
Just speculation on my part of course, so take it with a grain of salt, but it makes sense ot me.
Er.. You’re not missin much at all. Frankly, it’s a pretty useless utility. Better replaced by knowing a bit about where you put your files.
I think google search was just put out for a media push for google. Supporting Linux doesn’t get you nearly the media push as putting a almost-useful utility onto a Windows desktop.
If I were you, I’d be more mad at Apple for not porting iTunes….
The true Beagle is nothing more than Google Desktop Search on Windows. Not to mention it is poorly designed (I mean User Interface part etc.). I would better go for grep.
And your point is?, if Google dont do it for linux then we will have to do it ourselfs, we do that anyway so no loss.
Beagle poor UI, you must be joking right?, I mean how easy do you want it.
Actually I was not talking of ease but functionality. See Windows Search thouch 4 years older than Beagle its pretty well organised (ie when you disable the animated creature). Beagle’s Interface doesnt look like its something very great (though on Linux it is). It doesnt give that feel or quality what other competitors give in other OSs too (Mac OS X Tiger for eg.).
I watched a thing on writing code to make use of the search framework… It looked to me as if you could very well design a CLI utility in 10 minutes to use the search funtionality.
I forget where the video was, it’s been a while.
While Beagle is definitely still a late-alpha/early-beta release (crashes and resource hogging are something to be expected)
Crashes may be expected for an early version, but decent resource usage should be designed in from the start. Otherwise they’ll just have to rewrite the whole thing and fix the bugs all over again.
Crashes may be expected for an early version, but decent resource usage should be designed in from the start. Otherwise they’ll just have to rewrite the whole thing and fix the bugs all over again.
Well, part of the issues are with Mono as well, I believe. Mono is an important project, but needs to mature. Soon.
Beagle works well, though. For the most part.
In 2 years beagle’s team was unable to complete a decent
desktop search like google.
and yes,beagle is owned by novell employees
Your acting as though it’s a easy task, i’m sure Google ran into problems, but guess they have not the experites to make a Linux version or even know how opensource works!
Beagle works well, though. For the most part.
that’s why you never seen GDS/spotlight
I think the author somewhat misunderstands the capabilities of searching technology and how it can be applied to current hierarchical systems. First of all, any kind of non-hierarchical system for orginizing data will not work, period, whether it be relational, or just beagle-like.
The problem with relational/search-based file management systems isn’t that they’re bad at finding files (they’re great at this), but that they’re bad at showing higher-level relationships, and make it harder to visualize the overall structure.
Folders and directories are heavily misused, in my opinion. Specifically, deeply nested hierarchies are a big headache to deal with, and should be avoided.
I remember having a home directory just like the authors–it was full of junk files that really belonged somewhere, but I was too lazy to clean it up. In fact, I’d gather to say that I have an even larger home folder than the authors, with 13038 directories and 152175 files.
Then, I decided to follow the advice in this article ( http://evolvedoo.sourceforge.net/abstract/ ) and make my home directory my desktop. It’s incredibly easier, now, to manage all my files, and I actually know where things are. Better yet, I don’t even have to reorganize things anymore, they just fall into place.
Anyway, just wanted to point out that folders are very effective tools, they’re just used improperly. Relational/search-based systems (appear to) solve the immediate problem, but introduce a much worse one in doing so.
-bytecoder
Again, the KDE guys did it right. Although I’m a SUSE fan, and therefore have a default Beagle config, I’ve switched to Kat. Kat is a KDE desktop search application that is brand new, but already more stabile and better than Beagle.
Forget Beagle. Use Kat.
KAT site: http://rcappuccio.altervista.org/
I agree with ralph – your filesystem doesn’t have to be a mess for beagle to be useful. My number one use for it is when someone asks me a question, and I vaguely remember coming across the answer some time ago but I can’t remember when or where. It could be on IRC, on gaim, in my email or maybe on a website I visited. Now I could grep my xchat logs, then my gaim logs, then run an evolution search on my email and then hit google – or I could just type some relevant terms into beagle, which searches all of them at once and has a fairly good stab at ordering the results. Which is more useful, do you think?
Hmmm…
You’re the first person to make this sound useful to me…
Seriously, I’m not being sarcastic.
heh
. It really is useful that way. If I want to listen to a song…I don’t need beagle. If I want to open a document…I don’t need beagle. Don’t use it for anything like that, because I know where my data is. But if I don’t know _exactly_ what it is I’m looking for, yeah, it’s useful. Maybe I only use it once a day, but it’s enough to be worthwhile.
anonymous re kat – are you on crack? Seriously. Beagle has its problems, but it’s actually a useful and well-designed search tool. kat is the single worst designed (from an interface point of view) application I’ve ever seen. And going by the number of bug reports its inclusion in recent Mandriva betas has caused, I’d hardly call it stable. I wrote a blog entry on Kat’s inclusion in MDV 2006 default desktop, why I think this is a bad idea, what’s wrong with kat and why beagle’s better here:
http://www.happyassassin.net/2005/09/01/kat/
beagle ain’t perfect, not by a long shot. But I’d take it over kat any day.
Seriously. Beagle has its problems, but it’s actually a useful and well-designed search tool. kat is the single worst designed (from an interface point of view)
Well yer. They’ve got a rough and ready, not great, front-end to test it, but Kat is not a user interface, is it? It’s an underlying framework. Kat itself is not simply a user interface.
The link to the blog you’ve posted (well, it’s you) is written by someone who hasn’t the first clue what Kat actually is, because he is posting screenshots and whinging about them. That isn’t what Kat is. This person is also ranting and squirming that Mandriva is using what he perceives to be a KDE technology. He then wanders off talking about Gnome philosophy…..yada, yada, yada, and so on and so forth.
“Number three is that it’s a bad idea on principle to practically make it compulsory to have a flabby, non-essential application on everyone’s desktop – even users who obviously want a fast, thin setup, like people who install fluxbox or something.”
Well, what do you think Beagle is? It’s pink unicorn time again. Would Beagle installed by default be a bad thing I wonder?
And going by the number of bug reports its inclusion in recent Mandriva betas has caused, I’d hardly call it stable.
Wow. A totally brand, spanking new piece of software (or indeed, any piece of software) generating bug reports. Who would have thought that?! It’s going to be installed on Mandriva 2006, and uhmm, it isn’t 2006 yet.
btw, on the topic of other desktops being able to integrate search / find functions…
http://www.gnome.org/~fherrera/blog//searching
Don’t you mean “used it, abused it and reviewsed it?”
“Like some horrible RPM-induced circular dependency (sorry, Debian fan, that was a low blow) I seemed doomed to just repeat the process over and over.”
Shouldn’t he be apologizing to Red Hat fans?
“Like some horrible RPM-induced circular dependency (sorry, Debian fan, that was a low blow) I seemed doomed to just repeat the process over and over.”
Shouldn’t he be apologizing to Red Hat fans?
I didn’t get that, either.
-bytecoder
“Like some horrible RPM-induced circular dependency (sorry, Debian fan, that was a low blow) I seemed doomed to just repeat the process over and over.”
I believe he was identifying himself as a Debian fan. Read the sentence out loud and it seems to work.
Kat is not very usable, IMHO. When it first came out, it sounded like (from the author’s description) it was a fun weekend project quickly thrown together with little thought. Then, I installed several versions of the program, and it always felt like a quickly thrown together app. The backend is slow, and the frontend is very confusing.
Quick gripes: indexing takes half a day to two days and creates a 200MB – 2GB index; typing a search means waiting a minute or two for a result; the user must manually click “Create Archive”; the user must click “Open Archive” before starting searches (neither of these are very clear); and, index creation is a one-time, user-forced action (i.e., adding files to your directories requires a manual, lengthy index rebuilding).
This isn’t even a good prototype of any amazing backend search solution or amazing frontend interface. Beagle, OTOH, has a decent backend that actively updates the index, and has a usable frontend. At minimum, Beagle is a good prototype; but, it’s actually fairly usable now.
I think your comments are completely biased. You are evidently and undeniably a GTK zealot and a MONO-Microsoft believer. Let me tell you that Linux does not need that Micro$oft shit.
One thing is sure: Beagle will never be included in any serious distribution not financially bound to Novell (I’m talking about SUSE of course), because it has been made only to sponsor MONO.
If you read the comments about Beagle here and there, you will understand that only a few of GTK zealots like you succeeded in installing and actually using it (and I suspect that some of them are bluffing about that).
Talking about Kat, the infrastructure is maturing very quickly and the already excellent job done by its authors will be backed up by the KDE community in the next months to deliver a state-of-the-art search framework. The GUI of Kat has never been thought to be the default interface. It is only a catalog viewer and therefore not much attention has been payed by the authors to it.
Kat will be integrated so deeply in KDE that you will not even know that you are using it during your normal work. Its services will be available from every KDE application in an easy and coherent way.
This can only be accomplished in KDE because it is a really integrated framework.
This can only be accomplished in KDE because it is a really integrated framework.
Hah. That actually made me laugh a little.
Carry on.
(That’s not considered feeding the troll, is it?)
-bytecoder
“This can only be accomplished in KDE because it is a really integrated framework.”
Hah. That actually made me laugh a little.
Carry on.
You made me laugh quite a bit, because KDE quite clearly is a well integrated framework and is ready and waiting for technology like Kat. Please, do carry on laughing.
You made me laugh quite a bit, because KDE quite clearly is a well integrated framework and is ready and waiting for technology like Kat. Please, do carry on laughing.
I wasn’t laughing at that. I was laughing at his implications that it can be implemented on KDE and KDE only, as if there was some physical force preventing you from writing it for any other platform/library. He might as well be saying KDE was blessed by the gods–it would be about as true as his previous statement.
I wasn’t laughing at that.
Really? That’s what was quoted.
I was laughing at his implications that it can be implemented on KDE and KDE only, as if there was some physical force preventing you from writing it for any other platform/library.
He didn’t say that, although Kat does depend a lot on KDE libraries and technology for the level of integration they want. All he talked about was that it would be integrated deeply into KDE and touch every part of the system where you would need search/find technology in a way that is so seamless you won’t need to think about it – although that is for when Tenor does get rolling. I don’t know where you picked up the idea that he was implying something different.
Somehow when a Gnome/Mono application does that (or tries to do it) it’s the greatest thing since sliced bread and when KDE (or a KDE oriented bit of software) does it it’s somehow evil and the spawn of Satan.
In terms of integration I would probably expect no less than that level from Beagle within Gnome.
He might as well be saying KDE was blessed by the gods–it would be about as true as his previous statement.
Well you might as well say Mono is the greatest invention since the wheel and Beagle is the greatest thing since the motor car. That’s basically what you’re saying, so I don’t know how that is any different to an opposing view. Pink unicorn time again.
Beagle is already included in Mandriva contrib. I installed it by doing ‘urpmi beagle’. Admittedly I had to enable extended attributes for /home by editing /etc/fstab manually, which would be beyond a non-advanced user, but it was well documented in the Beagle wiki.
“Talking about Kat, the infrastructure is maturing very quickly…will be backed up…in the next months…will be integrated…will be available”
That all sounds great! If the backend matures properly – the database gets a lot smaller, a lot quicker to generate, and is automatically updated in response to filesystem changes – if the UI is improved and integrated into KDE, brilliant. You will then have a fantastic desktop search tool and I couldn’t be happier for you.
What you have _now_, though, is a borderline unusable, incredibly top heavy, badly engineered piece of crap. And it’s now that I’m talking about. Right now, Beagle is useful and usable (though still resource intensive and prone to memory leaks which mean you have to restart the backend every couple of days). Right now, Kat is essentially useless.
What you have _now_, though, is a borderline unusable, incredibly top heavy, badly engineered piece of crap.
I take it you’re talking about Beagle, because there’s no basis for saying that about Kat?
Besides, the goals of Kat’s integration into KDE is not to have a search tool but to provide a contextually accurate find system. There’s a huge difference.
Right now, Beagle is useful and usable (though still resource intensive and prone to memory leaks which mean you have to restart the backend every couple of days).
Wow. Considering Beagle has been around for quite a while and it is still essentially useless (I do not call the above usable by any stretch) I hold out more hope for Kat.
Maybe they should use proper development technology for Beagle as opposed to rewriting already decent software in C#?!
Has anyone tried Copernic or Googles program under WINE? I use WinZIP under WINE, and it works great.
You can grep through .pdf’s using pdftotext, and grep through .doc’s using antiword or catdoc. A little shell script which identifies a file as a .pdf or a .doc, and then using pdftotext or antiword or catdoc or libwpd (for WordPerfect files), before piping that output to grep should do the trick. Not exactly for newbies, but at least it’s smaller, and it works today.
C# is a IS0 standard
.NET C# is microsoftes implimentation
Mono is the Novell/OpenSource implimentation
Also,
when he says (sorry, Debian fan…
He isn’t saying sorry to Debian fans, he is saying he is one.
-voidlogic
“C# is a IS0 standard
.NET C# is microsoftes implimentation ”
The implementation is only available under RAND terms NOT compatible any of the free software licenses. So being a “Standard” doesnt mean shit here.
It means a lot actually. Being a standard means it’s open to reimplementation; which is what Mono’s c# is. This is why people are willing to write Mono apps.
And PS to OSnews… The Vonage ads are obnoxious. They often place incorrectly on Opera 8. Please, get a decent ad distributor or something..
The standard applies to the specification or programming interface. The license provisions of the Microsoft’s .NET implementation is irrelevant. Mono was only possible because the ECMA standard is open. The license provisions of the Mono implementation are all that matter to FOSS.
Yes under a RAND licence
and Rand means
Reasonably and Non Discriminatory but … Reasonably to me ? or other people ?
Given that Kat is only 3-4 months old while Beagle is someting like 15 months old, it’s not surprising it is little ahead. The real question is for how long.
Just getting really tired that Beagle gets all the attention. Kat is really the way to go : no dependency on Mono and KDE-based !
“That isn’t what Kat is.”
Really? It got installed when I did urpmi kat. It runs when I type ‘kat’. It’s got kat at the top of the window. If it walks like a kat, smells like a kat and miaows like a kat, I’d say it’s Kat. Now it may have a fantastic backend and wonderful plans for future KDE integration, but right now, that’s what you get when you download kat and run it. I don’t CARE if it’s a proof of concept or no-one cares or it’s going to be replaced by something made of solid gold in two weeks: THAT IS WHAT KAT IS RIGHT NOW.
As for my blog, you need to get better informed. I work for Mandriva, it’s basically an internal plea. And yes, Beagle installed by default would be a bad idea too (though better than Kat, minorly). And it’s being included specifically because it’s perceived (by our PR department…) to be ‘desktop search’. It’s being marketed as such in the PR planned for 2006. Oh, and 2006 is being released this month (or next month if it slips). It’s in version freeze, so whatever the wonderful developments to Kat that are just around the corner, they won’t be included.
Oh, and as for the other argument – the original poster used the word ‘only’, as if it were ONLY KDE which was such a wonderful framework that it could handle an integrated search (or find, or whatever) tool. I don’t think anyone’s saying KDE is _not_ a wonderful framework. They’re saying other desktops are good enough to handle an integrated tool as well.
Python, Ruby, and other scripting languages always will the toys for amateurs and not tools for professional developers. Lisp: it is not a new thing, and nobody use it except the emacs editor…
Python, Ruby, and other scripting languages always will the toys for amateurs and not tools for professional developers.
Except those languages are much more powerful and expressive than a braindead language like Java, and to a lesser extent C#
IMHO there is not any fundamental difference between procedural languages. If you have any problem, you can solve it with C#, Python, Pascal, C++ in similar way. The only biggest difference between C,C++, C#, Java, Pascal, etc and the scripting languages the dynamic typing. And IMHO this feature is more harmful then useful in bigger projects.
If you have any problem, you can solve it with C#, Python, Pascal, C++ in similar way.
Yeah, I could solve the problem in Assembly, but that doesn’t mean I should or I would want to.
The only biggest difference between C,C++, C#, Java, Pascal, etc and the scripting languages the dynamic typing
Not all “scripting” languages are dynamically typed. People throw out the word scripting without even really know what they’re talking about. Most Ruby or Python programming is not scripting.
And IMHO this feature is more harmful then useful in bigger projects.
Do you have data to back that up? The verbosity of Java (dumbed down C++) is probably more harmful than useful in bigger projects. Remember, people have to read the code and having 3x as much code because the language isn’t expressive doesn’t help large projects either.
Part of the reason that makes Ruby on Rails possible is because the language is highly dynamic with metaprogramming features.
“Remember, people have to read the code and having 3x as much code because the language isn’t expressive doesn’t help large projects either”
the logic of that statement implies that you need to read all code to understand the program. i belief that is clearly a false assumption. I even belief that declarations help you by having to read less code . for example to find what kind of object is returned by a python method you need to study the code. (a commend is not going to help you as much as that is not as reliable as it is not checked by the compiler).
“Yeah, I could solve the problem in Assembly, but that doesn’t mean I should or I would want to.”
Yes, but the gap between the assembly or any modern OOP programming language bigger then the programming languages and scripting languages.
“Not all “scripting” languages are dynamically typed”
Yes, but all dynamically typed languages are scripting languages.
“Most Ruby or Python programming is not scripting”
You can create *nix-like “OS” in javascript. But IMHO javscript not a low level language.
“Do you have data to back that up?”
I created many programs with dynamic typing and static typing systems, I know the differ. Until you must read/write only your code (or any good programmer’s code and the code is well commented), the code is not too big, and the time is not too important the dynamic typing is a good thing…
KAT’s GUI is known to be bad, even the developers admit that 😉 It was hacked up in the early days of the project, and not changed much since then. But a better interface and better KDE integration are planned for the next release.
Apart from that, KAT will be cooperate with the Tenor context linkage engine, which will just rock once its there. It will offer totally new concepts for linking content on the desktop and browsing it by context. Beagle isn’t rocket science, its just yet another desktop search engine and quite boring from the conceptional point of view.
Admittedly, it “just works” (TM) while Tenor is still vapourware.
>I wasn’t laughing at that. I was laughing at his
>implications that it can be implemented on KDE and
>KDE only, as if there was some physical force
>preventing you from writing it for any other
>platform/library.
Well, indexing like beagle and kat do it now is no rocket science and actually boring and limited. A context engine like Tenor (which is in conceptual stage) can link information in the background (save a mail attachment, Tenor saves the link mail->file. The UI of KDE can then utilize the link and e.g. show the sender of the mail when browsing the file).
This cant be done from the outside, but needs integration directly into the app, not some shitty indexer plugin that just reads static data. Of course this is also possible in gnome, but KDE is more integrated and has a more “centralized” community (which has both good and bad aspects, here its an advantage).
are the future. People don’t want to admit this, but just look at IntelliJ, things are heading towards domain specific languages and dev environments (called the Language Workbench by Fowler) that just adapt, IntelliJ is a glimpse into the future of this and static type is needed for good code analysis and supporting semantic level automated tasks, like advanced refactoring and so on.
I’ve sat here and read through these comments whilst waiting for a big job to finish. Otherwise I wouldn’t bother. This discussion just goes to show how time wasting, bickering, flaming clueless trolls have infested the FOSS community.
As an end-user, _who cares_ what language it’s written in? You want speed, usability, integration, etc. The language is the concern of the developer, no-one else. Stick to some uninformed opinion about what language you want your software written in and you’re cutting off your nose to spite your face.
What’s with all this GNOME vs. KDE nonsense too? As an end user, again, we want integration. Best tool for the job. The concept of a distinct ‘desktop’ is a terribly deprecated notion. Too many lusers are getting attached to something they have no reason to be zealous about: flaming in such ridiculous threads as this does not count as making a contribution or being a part of a community.