All my friends who deal with professional publishing say the same. Everyone in industry tried GIMP - it's just not up to the job. Or is there any one here silly enough who would cash out $600 for Photoshop (aka PS) when there is gratis alternative? All I can say is that some companies bye for most people cheaper Corel PhotoPaint - but for viewing purpses only. Photoshop with its price tag is reserved for designers only.
PS pros prefer PS. For users who need e.g. QC/QA designer's work GIMP's GUI is bit too frustrating, but Corel's PhotoPaint is Okay.
Most my friends designers have jobs related to printing and/or polygraphy. And all of them take part time jobs for web designe and/or development. Guess what kind of tool they use? Free GIMP? - or Photoshop they already have and know?
For any design related job, polygraphy is sacred art. W/o understadning inner working of polygraphy and publishing one would never become good designer and would never master Photoshop. That's what most of my friends are saying. (Artistic taste helps a lot too.) Advertisement market is hugest consumer of DSP (Desktop Publishing) - Internet ad market is growing - but still just little fraction of ad market of newspapers. And most of that market is served by PS.
In fact, I think comparision of GIMP and Photoshop is rather silly. Photoshop is built from ground zero for preparing raster graphics for publishing. GIMP has no target market. (Well, it has no market at all - it's not sold (-; ) Due to that difference, GIMP cannot mature into good publising tool: there is no strong driving force before/behind it. Adobe has some obligations before its customers - and it wisely chooses to listen to them. GIMP doesn't experience that kind of user pressure.
It doesn't mean GIMP is bad. It's great tool for preparing content for Web. Most of Linux GUI is made in GIMP. It also has excelent scripting. In fact, from what I know, GIMP's scripting is the most wanted feature by most of my PS-pro desginer friends. (One my friend learned subset of Perl and installed Linux specifically for GIMP/Perl tasks.) Photoshop has something to learn from GIMP too.
It definitely sounds good. At least it removes another 'contra' from long list of IA-32/AMD64 and PowerPC differences. As many of you have known for some time people who run Linux on PPC enjoyed free ride with Mac-On-Linux project. Check http://maconlinux.org/
On side note, after seeing how easy virtualization can happen with Open Source kernels - e.g. User Mode Linux, Xen, Plex, coLinux, etc - me keeps wandering why M$ haven't done that with WinNT kernels. There are only few true obstacles in x86 "architecture" which prevent effective virtualization - VMware is solving all of them at very high level and of course tried in past to charge premium for that. Xen modifies kernel so that overhead of virtualization is negligible - it's not another computer emulator, it's just kernel running as a ordinary OS process. (Anyway, user tasks see computer only as it is reflected by kernel and device drivers (-: )
I know M$ likes only good cash cows (like M$ Office franchise) but as OS kernel concerned, the modifications to allow it to run in virtual machine are truly not that big. Check-out the coLinux - it's neat. http://wiki.colinux.org/cgi-bin/ConvertingDistribu tions
P.S. Or is it what M$ Windows Advanced Server for?
Re:It really is true...
on
Beyond Java
·
· Score: 1
As grand parent have put it: "Compared even to something like C++, Java's syntax is cumbersome and full of unnecessary repetition."
After several years of experience I have found that over-complex syntax has more or less direct performance implication too.
As a huge Perl fan & pro, once I have read on how perl optimizes list/map operations. To put it simply, since Perl syntax tends to be short and all list/map operators are part of language itself, Perl optimizer has list of standard optimization patterns for many use cases.
Compare to Java. (My experience with Java is quite out-dated and Thanks God I'm finished with Java in 1.2/Java2 times.) I needed to implement simple class, analog of grep: one list comes in, two lists comes out. Code to do that took about 20 lines. Compare to Perl's one line with grep or map. It's obvious that Perl has much much greater chance to optimize complex single line operation, rather than Java compiler (how sophisticated it is) tinkering with 20 lines of code.
I find most of the time people being more ignorant about Java performance, since it is very hard to track were performance was lost and how it can be recovered. Java applications can be fast. But as number of classes and objects grow, absences of decent memory manager becomes obvious obstacle.
Everything proper has compromise in it. As modern memory-hunger algorithms concerned, Java failed to realize proper compromise.
Check C++: you write as much code as in Java, but C++ run-time doesn't have to bother with memory management. (And in most applications memory management can be optmized - so in most of the cases there is even no programming problems.)
Take a look at Python or Perl: you write many time less code as you do in Java, and interpreter run-time (due to Python/Perl rich operator set) doesn't have to guess what code does and better way to optimize it. (In many situations, of course, most optimizations of Perl/Python are specifically targeted at avoiding triggering memory allocation. Memory allocation is always expensive.)
And now compare to Java: you write quite much code. And Java compiler still has hard time trying to figure out what the code is doing and how to optimize it better. Lose-lose situation. And you cannot avoid triggering memory management. C/C++ you can put object on stack: in Java every object is in fact pointer to object in global memory pool. Lose-lose-lose situation. Developer has to fight both compiler and run-time libraries. Run-time libraries has to be ready for lots of wierdness from a developer - he would try to optimize. Compiler has to be ready for even more weirdness from both run-time library and user code... Like a snowball - the pains grow like snowball with every release of Java. (At least that's my impression.)
Brrr... Was there for short time. Thanks God not there anymore.
It allows to compare how M$ Windows executable under Wine performs relatively to native Linux executable and to native Wind0ze executable under Wind0ze.
> "Clusters are only really good for embarassingly parallel problems."
You are overall sort of right about that. It's just science isn't standing still and most of new algorithms are specifically invented to be parallelized.
Most problems of physics are solved with matrices. And matrices are of course are easy to parallelize. And physics - is the only who are buying most of big iron anyway.
Nowadays, most of the weather prediction tasks, astronomical tasks, optical tasks, micromeasurements tasks are also optimzed for clusters - not big iron. It's not about top performace - it's about price/performance ratio. For the same money people can buy cluster with e.g. 10 times more raw performance to run unoptimal (e.g. 2-3 times slower) algorithms - but task are done quicker. And cheaper. Yeah, clusters have higher latencies - but they are still dominated by batch jobs, not interactive jobs. Big Iron has better interconnects - but the redundant interconnects take lion share of such system costs.
In fact, the main reason why this have happened (clusters took over big iron) is the RAM prices. In my versity times (early-mid 90s), we all were occupied with shared memory problem: RAM was very expensive. Now people go to general store, pick several 1GB nics, pick several GBs of RAM, and pay a nickel for all that.
Ask anyone in Computer Science now, everyone started throwing RAM at latency problems of clusters. It does look bad on paper and in theory. But in practice it just works.
P.S. On-topic. IA-64 has great performance. But again, on price/performance scale it loses immediately to Intel's own Xeons and AMD's Opterons. Intel constantly refuses to amend its Itanic focus from features to focus on more affordable prices. The story line was quite well covered by The Register. Check posted links.
As Linus have stated that depends on particular source code file.
GPLv2 is compatible with GPLv3 and vice versa.
You can have the project with files under both licenses in the repository. Both GPLs have the same requirements on redistribution - and that what matters in case of Linux kernel.
As soon as you will start changing something - then you might hit the problem of incompatibility of GPLv2/GPLv3. But as FLOSS developers (like I am, or Linux & KDE crowds) are concerned - they are equal. It's not that I'm going to release DRM covered by pile of patents. Nor I'm going to use such stuff.
GPL is getting adopted to wider range of products. So FSF now merely explains all fine points one could have derived from GPLv2. Before GPLv3, some things (like DRM) were complicated to put under GPL. GPLv3 makes another step and makes such things impossible.
GPL is more about ideology of development - not about rights and obligations. It just uses legal speak - which is rights and obligations - to describe that ideology. IMHO.
Then they are all evil terrorist threatening U.S. liberty!!!
Lopata.
P.S. The major problem of the iTunes Store (and any subscription service for that matter) is that you have to have good internet connection. And still, video of good enough quality takes lots of space. (Think DivX or MPEG4/AVC: 10 minutes ~ 100MB or more). Even if Apple would start selling h.264 full size movie downloads - who would want to pay money for ability to kill connection for a whole day or more? With all the DRMing stuff that would be too burdensome to customer. So for now Apple decided to go with something they can fit into appropriate download size.
But on other side, this is the only option at moment. Quality streaming had proven to be undoable on internet: average (read: cheap) home connection (infamous last mile) has way too high latencies.
I have bystander knowledge of IBM only. All I can say this guys have too much experience.
They quite good at separating "good product" from "cash cow" lines. What you describe is sort'a pet projects for investors. Normally this is products which look great on paper - but in reality make no sense. Investors of course care more about what's on paper. We as customers care more about product.
As an example, take one of my projects. (*Not* for IBM.) The project consisted of two parts: big ugly brick under table and external controller (a-la joystick with screen). Excluding my manager, there were about 5 more managers who were "tracking" the project. Guess which part of it they were interested? Of course joystick with screen! My manager used that fact to distract people from main system unit and give me time and place to do my job. The guy who was on the joystick part had less luck (but quite much fun): day long meetings every week about how joystick movement feels, some sales coming constantly and arguing long about where company logo has to be, etc. You can imaging. Secret was that joystick was ready about the time hardware for main system unit was ready - iow it was finished already when all meetings took part. But since it was unusable w/o system unit - it was considered part of single project, rather than separate project. We were ready on dead-line. What we did one way or another highlighted that all those meetings were waste of time. "Nevermind" as my manager used to comment on that.
That has taught me that in some companies managers in fact has to do lots of work to allow mere developers to do the job. When you look inside - it might look like scam. But as long as job gets done and company has something to sell - that's okay. In big companies it is just too much interested parties - some-one has to keep them at bay, so that normal work can go on.
I wonder if that would hint Apple to licesning Windows Media for Mac OS from M$.
P.S. Thou, I think that's impossible, due to Apple's policies on QuickTime: ability to play files created with older version of QucikTime. M$ likes to drop stuff as they wish. Apple somewhat cares more about customers' experience.
P.P.S. I keep that WMP for Mac on my iBook - sometimes I'm getting WMV files from Wind0ze people. But, Thanks God, many now came back to DivX/XViD with MP3/AAC packaged as AVI or OGM. So most of the time it works now;-)
P.P.P.S. I beleive, WMP for Mac always had the "Beta" moniker. Thou I think we can hardly call that "being released". Anyway, VLC is better. It would be good, if one day it will support WMV9 - but otherwise it is not something that I miss much. Anime fansub groups release DivX which plays with no problems by all players;-)
P.P.P.P.S. "Our focus really is in delivering the best experience to Windows customers." Then M$ has to drop that crap called Windows Media and license and bundle DivX. That's my best experience on Wind0ze.
Why people in wealthy countries never think of moving into developing countries?
Thou I haven't tried that by myself;-)
Freinds of mine - married couple of Uni teachers - were working in North Africa (Marocco & Alger) for about 10 years. When they came back I only hear them complaining how much they earn here but can buy literally nothing. Situation in Africa was different: they were earning little, but most of things costed next to nothing. In Africa they were top - here they are just average. (They came back because both wanted to have children in homeland.)
What just reminds me about how overvalued, overpriced developed countries became with time.
As many have already speculated, HDTV quality is high enough to be ripped from analogue signal and converted back to digital. Anyway, DivX/XViD image size is normally scaled down, compared to DVD resolution - iow we are already used to lower-than-DVD quality. And it is Okay;-)
Case 1: you buy CD, rip it and share mp3s on net. Case 2: you connect to net and download some mp3s.
In first case you are criminal, since you knew that CD content is copyrighted, disregarded that and exercised the right you were not given by copyright law.
In second case you are innocent, since you have no clue about who is owning the content. And it is duty of copyright owner to explain that to you and ask to stop illegal distribution.
Modern P2P networks have merged both cases together - since everyone is uploader and downloader at the same time.
Thou again: to prove your wrongdoing in first case, the copyrighted work must be downloaded completely from you - but not someone else. If you have downloaded file from 100 hosts - none of the 100 people behind the ip addresses has uploaded complete work. RIAA not yet tried to crack on BT or eMule networks: before that we wouldn't know what U.S. legal system thinks about such kind of distribution.
MySQL is fast but it has sacrificed many features for the performance. Big DBs they might be slower - but over all they do provide more features and (very importantly) robust performance.
I have worked with MySQL, MS SQL & Sybase.
Sybase of Linux left the best impression overall. I haven't digged much Sybase on Windows - so I cannot tell. It just worked. Server crashed (or was just accidently unplugged) - it was getting up with no fuss. Scripting in the beggining was very confusing but few day of learning payed back much.
MS SQL Server pares on features but performance-wise needs a dedicated server - even if load is low server is barely usable for anything else. We tried to run IIS and MS SQL Server on single machine but later abandoned that idea. After Sybase complete control from command line, MS's love to silly configuration GUIs was bit irritating. (Especially when you still end-up firing regedit) Later on we discovered that MS SQL is very unhappy because of manual tuning applied - going back to defaults helped.
MySQL on Linux was pure fun. You can insert 1-2k rows in table per second using normal INSERT. (MS SQL/Sybase could both insert few hundred rows per second - for large operations it is recommended to use backup/restore facility) But then after say 100k records fun was beginning - MySQL was driving complete Linux box into eternal swap nirvana trying to index inserted data (which of course didn't fit the RAM). Who is guilty? Linux which over last 4 stable incarnations only now ca barely effectively handle files bigger than RAM? Or MySQL which wasn't expecting that someone will try to create table for 100k rows? I do not know. Then you did something like that "SELECT * FROM a WHERE a.a=1 or a.a == 3" over that 100k table - and you are going to drink some coffee - MySQL would need 10-20 minutes to finish the query. Why? - "OR" operation is not optimized and not a priority for developers. MySQL development was always very focused: if you can achieve something easier by hack elsewhere - do hack elsewhere. That's the response I got from Monty on my OR'd select problems: don't do selects with OR, instead do OR manually in your application. We later had to port complete stack from MySQL/Perl over to MS SQL/ColdFusion. It wasn't nice - I still have recurring nightmares of ColdFusuion4 scripts - but it worked Okay.
Well that was situation with MySQL back 5 years ago. For big systems we were using MS SQL/Windows & Sybase/Linux - for small stuff, especially for VERIO customers, MySQL was standard. You definitly do not need something big like Oracle for 3-5 tables barely exceeding 10k records - and that conf covers about 75% of private B2B portals.
What they want? I'll tell you what they want. Subscription service you cannot unsubscribe from with Britney Spears on every channel all the time. They do nothing and you pay them money. All the time. And you have no escape. Even if you do not watch - you still pay. That's what they want.
P2P file sharing and Internet gave people that "always on" feeling. But people refuse to pay fees **AA wants to charge them. **AA still lives in "on demand" times (of CD sales) as opposed to "always on."
Right now legal "always on", for example, is music TV channel, where you pay your subscription by watching lengthy ads. They have screwed radios and TV stations with lengty contract on how they can use the content. Now **AA want something like that for Internet. But Internet, created by Unix hackers, just lack proper ways for them to regulate how and what you are listening.
iTunes is sort'a compromise of the "always on" and "on demand" systems. But subscription services are what **AA truly commiting to: sustainable revenue stream. You know, artists are not stable, it's hard to do business with them.
When company found to be a monopoly, regulation body would be created. On the committee would be people from the company, its competitors and independent professionals from application field.
Such committee will oversee all actions of monopolist for some time (depends of the court order). Competitors can submit their complain and wishes - committee will decide on them what to do. Such committee can effectively over time put more restrictions on M$ - but only in line with court decision.
Deutsche Telekom (DT) of Germany is monopolist and regulated in that way. Court ordered them to open part of network for smaller ISPs thru number of connection points. Later on smaller ISPs have complained that DT has made the connection points practicly unusable to provide any good level of service to customers. In response, regulation body has investigated and confirmed that to be the case and has ordered DT to provide more connection points. After original court decision there were about 10 independent ISP here. Now - after regulation body intervention - about 50.
IOW, If M$ will submit incorrect or out-dated documentation, qualified people would notice that immediately. They will be involved in that process directly and report M$ incompliance.
How true. To the last word. Except that Linux use requires IQ130+ - my IQ is about of level of collective intellegence of slashdot population.
Seeing overtime how corrupt Europe is make me very sad. And creation of EU has just created many more cash cows the bureaucracy can milk - just recall money laundering scandal with EuroStat which 3 years later now is still nowhere to any resolution.
I think have M$ played the proper political tune, it could have delayed the case as long as it wanted. But M$ as true american company tried to muscle too much, what didn't frightened local politicians. Instead they have seen opportunities to make good careers themselves. M$ didn't get the difference between U.S. and Europe. It didn't expected that some will choose career later over money now.
As things now, M$ will have to throw some bone to the angry dog. Probably it will be considerate enough to make a friends with its owner.
P.S. On other side that might be a delaying game of M$. Delay long enough til when the APIs in question would become obsolete enough. Or even better til when Windows Server line will be dropped in favor of Vista Server/whatever but not "Windows Server". And they are off of the hook.
"""I just had an interesting thought that your comment about HTML+CSS prompted.""" """What if the presentation is the content?"""
Well this is why XML's style sheets are XML - XSL. And even more. Transformation of XML styles described in XML too. And they are standartized too - XSLT.
So you can transform style. And you can transform transformation. XML has enough rope for everyone.
And true American parents will be able to protect their children from the filth of The Real World: for example all scenes of violence will be automatically replaced by big red [CENSORED] banner right in the kids' brains!
Matrix! please have me!!!
I think RIAA/MPAA must build another country: with 100% secure content distribution, with 100% managed internet, with total protection of all those precious intellectual propery rights. And distribute precious content of theirs only there. Then they could run commercials of this land trying to persuade people to move in: "Joe Six Pack(TM) moved to WonderLand(TM) and patented letter A(TM)(P) - and now(RM)(pat. pend.) he is billionaire(TM)(P)! Now he has enough money to buy s/h TV set!!! WonderLand - land of new opportunities(RM)(TM)(C)(P)!!!!! [This ad(TM)(C)(P) in part brought to you by Joe Six Pack Eterprises(TM)(pat. pend.), exclusive(TM)(pat. pend.) licensor(TM)(P)(pat. pend.) of latin(C)(P)(RM)(TM)(pat. pend.) alphabet(TM) in WonderLand(TM)(pat. pend.)(RM)(pat. pend.)(TM)(RM)(pat. pend.). Reg-No: 8000-235435248-2343234. DRM-Id: 4334242-234234-2342342. Watermark-Id: 4343423456524234. MD5: 2323234542342424234234234. CRC32: 12123432. By watching this ad you give up all your rights in favor of Joe Six Pack Enterprises of WonderLand. FBI Warning: For illegal use of the ad, you might be sentenced for up to 50 years of life in WonderLand.]"
One was giving short insight into ODF v. M$' DOCX. (*)
ODF is flow based a-la HTML. DOCX is record based a-la files generated by Write (.wri) & WinWord' text changes stream.
Application of styles is very different. Even if conversion of text can be made, conversion of styles is almost impossible. In DOCX styles are more or less inlined - ODF was redesigned by OASIS with styles to be more like HTML+CSS.
Basicly, M$'s concept boils down to "anything can occur anywhere in document". ODF hence standard is more strict.
Additionally M$ has special support for ActiveX: embedded objects will be stored as binary dump in middle of XML documents. (E.g. all pictures and files inserted from outside in M$ universe are ActiveX objects.) I'm not sure who ODF pares with embedded content, I can only hope OASIS - unlike M$ - have put XML to good use. After all, unlike M$, ODF includes vector and raster graphics too.
That's actually why most pirated DVDs and movies have that warning/copyright info deleted. (*)
I normally recognize pirated stuff from original by presence of the overlong legal statements. True pirates normally do not reproduce it.
So if you have bought pirated DVD, you might see no warnings whatsoever. And copyright statement could be (intentionally made) incorrect too. I'm not sure how court would regard distributing copies made off pirated copy. Technically (if you stopped distributing upon noitice from copyright holder) you have seen no copyright statement so you are not guilty. Guitly would be the guy who sold you the original pirated copy.
(*) This is where actually DRM-like crap started creeping into DVD standard: MPAA wanted a "cannot skip" flag for particular section of video. Ever tryed to skip that green/red warning? - most DVD players will err with "Not Permitted" message. (In the time, everyone was afraid that cheap retailer might use that for advertisments. Thanks God that didn't happened.)
I'm voting for VIM. (Just for record: I used vi once on Solaris - it's pain. So please please please all those Linux users do not mess vim with vi. If you have used only Linux that would mean you never really seen horrors of vi - all of OSs of GNU/Linux fame use vim instead of vi.)
I've seen people used to use emacs. But I've never seen anyone who really knows what is going on inside. Or truely understands how some things work. People just got used to Emacs - not understood it. Emacs is like religion - you cannot understand it by definition.
I like VIM. It integrates easily with Unix tools and shell.
In contrast, Emacs integrates with nothing - it has replacement for all the tool normally one would expect to be implemented on OS level. Emacs is great tool, but learning it with all its modes and buffers could fill one's full-time position.
GNU - GNU's Not Unix, and Emacs is the confirmation. Try (re)configuring VIM and then try to configure Emacs. Emacs's init.el/friends always remind of sendmail.cf.
VIM doesn't have all the bells and whistles - but it does one thing and does it well. Emacs IMHO does many things - but does them averagely. After several years of unsuccecful tries to adopt almighty Emacs, I gave up on it. I know VIM and Bash - and can already do my job. With Emacs, regardless how much you already know Emacs is always ready to provide you with many new problems and obstacles. (Once I have tried to teach Emacs to use tab instead of spaces. It's very long and sad tale with no happy-end.)
VIM is simple and only hard thing to learn there is Insert v. Control mode. Once you got that, you can work already. I do not like it much - but I definitely like it more than cascaded N time nested M character long shortcuts of Emacs. VIM's simple shortcut structure also IMHO friendlier to touch-typists.
From programmer's POV, CC-mode is definitly superior to what VIM has to offer. But on other side, if something in CC mode doesn't fit your needs you might end-up killing two weeks to figuring out how to change that. VIM was hard to learn to use effectively. But once you learned it - due to its easy integration with OS - you find that you can make tasks you did before in shell better/faster/easier. (You can call Perl from VIM - or you can call VIM from Perl - choice is all yours.) It doesn't nullify your experience with Unix - like Emacs does - but rather magnifies it. IMHO.
Sony Classical is one of the biggest classical recording & CD publishing companies. If you will look into classical guitar (for example) you will find that about 90% of artists are published by Sony Classical.
I already hate SONY - with just this one little exception...
P.S. I recently bought a pile of CDs from Sony Classical, expecting to find a lot of (CP) markers, DRM toolkits, root-kits, thousand clause EULAs, etc. To my disappointment there were none included, even with SACDs.
I've been on one exhibition where some stats were drawn.
1st. Year 2003. In my application field, in North America (U.S. & Canada) ~650 start-up were founded. Europe? - 25!
2nd. My company - start-up - has made on exhibit to top 10 european start-ups. How? BY DEFAULT! We haven't yet managed to produce or sell something! It's just there were no other start-ups to compete against.
I'm living in Germany last 4 years - I yet to seen any progress at all. My poor home country Belarus - classified as "poor", "third-world" & "dictatorship" - sees more investments in development than "rich" Germany.
"Stagnation and protectionism" are two words I can use to describe local social and political systems.
And SAP actually is traditional German business working on traditional German principles. IOW. If complete idiot was hired - he will never be fired. (Competence of personel in Germany is really last thing anyone cares about. I'm working for third company and nobody - except staffers - ever looked into my CV.) All bugs are there to stay, since it can break numerous customer applications. 'Customer feedback' is something mythical, non-existent and ignored. Everything what have workaround is considered to be not bug, but feature. Ergonomics (it's over all German) doesn't ring any bells. Thick unpenetratable wall of management, secretaries, sales, service peresonel effectively shields any knowlegeable engineer from ever communicating with customers. And so on. I worked for similar company for some time.
Slashdot Mirror
I dunno why your post was marked flaimbait.
All my friends who deal with professional publishing say the same. Everyone in industry tried GIMP - it's just not up to the job. Or is there any one here silly enough who would cash out $600 for Photoshop (aka PS) when there is gratis alternative? All I can say is that some companies bye for most people cheaper Corel PhotoPaint - but for viewing purpses only. Photoshop with its price tag is reserved for designers only.
PS pros prefer PS. For users who need e.g. QC/QA designer's work GIMP's GUI is bit too frustrating, but Corel's PhotoPaint is Okay.
Most my friends designers have jobs related to printing and/or polygraphy. And all of them take part time jobs for web designe and/or development. Guess what kind of tool they use? Free GIMP? - or Photoshop they already have and know?
For any design related job, polygraphy is sacred art. W/o understadning inner working of polygraphy and publishing one would never become good designer and would never master Photoshop. That's what most of my friends are saying. (Artistic taste helps a lot too.) Advertisement market is hugest consumer of DSP (Desktop Publishing) - Internet ad market is growing - but still just little fraction of ad market of newspapers. And most of that market is served by PS.
In fact, I think comparision of GIMP and Photoshop is rather silly. Photoshop is built from ground zero for preparing raster graphics for publishing. GIMP has no target market. (Well, it has no market at all - it's not sold (-; ) Due to that difference, GIMP cannot mature into good publising tool: there is no strong driving force before/behind it. Adobe has some obligations before its customers - and it wisely chooses to listen to them. GIMP doesn't experience that kind of user pressure.
It doesn't mean GIMP is bad. It's great tool for preparing content for Web. Most of Linux GUI is made in GIMP. It also has excelent scripting. In fact, from what I know, GIMP's scripting is the most wanted feature by most of my PS-pro desginer friends. (One my friend learned subset of Perl and installed Linux specifically for GIMP/Perl tasks.) Photoshop has something to learn from GIMP too.
It definitely sounds good. At least it removes another 'contra' from long list of IA-32/AMD64 and PowerPC differences. As many of you have known for some time people who run Linux on PPC enjoyed free ride with Mac-On-Linux project. Check http://maconlinux.org/
u tions
On side note, after seeing how easy virtualization can happen with Open Source kernels - e.g. User Mode Linux, Xen, Plex, coLinux, etc - me keeps wandering why M$ haven't done that with WinNT kernels. There are only few true obstacles in x86 "architecture" which prevent effective virtualization - VMware is solving all of them at very high level and of course tried in past to charge premium for that. Xen modifies kernel so that overhead of virtualization is negligible - it's not another computer emulator, it's just kernel running as a ordinary OS process. (Anyway, user tasks see computer only as it is reflected by kernel and device drivers (-: )
I know M$ likes only good cash cows (like M$ Office franchise) but as OS kernel concerned, the modifications to allow it to run in virtual machine are truly not that big. Check-out the coLinux - it's neat. http://wiki.colinux.org/cgi-bin/ConvertingDistrib
P.S. Or is it what M$ Windows Advanced Server for?
As grand parent have put it: "Compared even to something like C++, Java's syntax is cumbersome and full of unnecessary repetition."
After several years of experience I have found that over-complex syntax has more or less direct performance implication too.
As a huge Perl fan & pro, once I have read on how perl optimizes list/map operations. To put it simply, since Perl syntax tends to be short and all list/map operators are part of language itself, Perl optimizer has list of standard optimization patterns for many use cases.
Compare to Java. (My experience with Java is quite out-dated and Thanks God I'm finished with Java in 1.2/Java2 times.) I needed to implement simple class, analog of grep: one list comes in, two lists comes out. Code to do that took about 20 lines. Compare to Perl's one line with grep or map. It's obvious that Perl has much much greater chance to optimize complex single line operation, rather than Java compiler (how sophisticated it is) tinkering with 20 lines of code.
I find most of the time people being more ignorant about Java performance, since it is very hard to track were performance was lost and how it can be recovered. Java applications can be fast. But as number of classes and objects grow, absences of decent memory manager becomes obvious obstacle.
Everything proper has compromise in it. As modern memory-hunger algorithms concerned, Java failed to realize proper compromise.
Check C++: you write as much code as in Java, but C++ run-time doesn't have to bother with memory management. (And in most applications memory management can be optmized - so in most of the cases there is even no programming problems.)
Take a look at Python or Perl: you write many time less code as you do in Java, and interpreter run-time (due to Python/Perl rich operator set) doesn't have to guess what code does and better way to optimize it. (In many situations, of course, most optimizations of Perl/Python are specifically targeted at avoiding triggering memory allocation. Memory allocation is always expensive.)
And now compare to Java: you write quite much code. And Java compiler still has hard time trying to figure out what the code is doing and how to optimize it better. Lose-lose situation. And you cannot avoid triggering memory management. C/C++ you can put object on stack: in Java every object is in fact pointer to object in global memory pool. Lose-lose-lose situation. Developer has to fight both compiler and run-time libraries. Run-time libraries has to be ready for lots of wierdness from a developer - he would try to optimize. Compiler has to be ready for even more weirdness from both run-time library and user code... Like a snowball - the pains grow like snowball with every release of Java. (At least that's my impression.)
Brrr... Was there for short time. Thanks God not there anymore.
That's why it is developer test.
It allows to compare how M$ Windows executable under Wine performs relatively to native Linux executable and to native Wind0ze executable under Wind0ze.
> "Clusters are only really good for embarassingly parallel problems."
You are overall sort of right about that. It's just science isn't standing still and most of new algorithms are specifically invented to be parallelized.
Most problems of physics are solved with matrices. And matrices are of course are easy to parallelize. And physics - is the only who are buying most of big iron anyway.
Nowadays, most of the weather prediction tasks, astronomical tasks, optical tasks, micromeasurements tasks are also optimzed for clusters - not big iron. It's not about top performace - it's about price/performance ratio. For the same money people can buy cluster with e.g. 10 times more raw performance to run unoptimal (e.g. 2-3 times slower) algorithms - but task are done quicker. And cheaper. Yeah, clusters have higher latencies - but they are still dominated by batch jobs, not interactive jobs. Big Iron has better interconnects - but the redundant interconnects take lion share of such system costs.
In fact, the main reason why this have happened (clusters took over big iron) is the RAM prices. In my versity times (early-mid 90s), we all were occupied with shared memory problem: RAM was very expensive. Now people go to general store, pick several 1GB nics, pick several GBs of RAM, and pay a nickel for all that.
Ask anyone in Computer Science now, everyone started throwing RAM at latency problems of clusters. It does look bad on paper and in theory. But in practice it just works.
P.S. On-topic. IA-64 has great performance. But again, on price/performance scale it loses immediately to Intel's own Xeons and AMD's Opterons. Intel constantly refuses to amend its Itanic focus from features to focus on more affordable prices. The story line was quite well covered by The Register. Check posted links.
As Linus have stated that depends on particular source code file.
GPLv2 is compatible with GPLv3 and vice versa.
You can have the project with files under both licenses in the repository. Both GPLs have the same requirements on redistribution - and that what matters in case of Linux kernel.
As soon as you will start changing something - then you might hit the problem of incompatibility of GPLv2/GPLv3. But as FLOSS developers (like I am, or Linux & KDE crowds) are concerned - they are equal. It's not that I'm going to release DRM covered by pile of patents. Nor I'm going to use such stuff.
GPL is getting adopted to wider range of products. So FSF now merely explains all fine points one could have derived from GPLv2. Before GPLv3, some things (like DRM) were complicated to put under GPL. GPLv3 makes another step and makes such things impossible.
GPL is more about ideology of development - not about rights and obligations. It just uses legal speak - which is rights and obligations - to describe that ideology. IMHO.
Lopata.
P.S. The major problem of the iTunes Store (and any subscription service for that matter) is that you have to have good internet connection. And still, video of good enough quality takes lots of space. (Think DivX or MPEG4/AVC: 10 minutes ~ 100MB or more). Even if Apple would start selling h.264 full size movie downloads - who would want to pay money for ability to kill connection for a whole day or more? With all the DRMing stuff that would be too burdensome to customer. So for now Apple decided to go with something they can fit into appropriate download size.
But on other side, this is the only option at moment. Quality streaming had proven to be undoable on internet: average (read: cheap) home connection (infamous last mile) has way too high latencies.
I have bystander knowledge of IBM only. All I can say this guys have too much experience.
They quite good at separating "good product" from "cash cow" lines. What you describe is sort'a pet projects for investors. Normally this is products which look great on paper - but in reality make no sense. Investors of course care more about what's on paper. We as customers care more about product.
As an example, take one of my projects. (*Not* for IBM.) The project consisted of two parts: big ugly brick under table and external controller (a-la joystick with screen). Excluding my manager, there were about 5 more managers who were "tracking" the project. Guess which part of it they were interested? Of course joystick with screen! My manager used that fact to distract people from main system unit and give me time and place to do my job. The guy who was on the joystick part had less luck (but quite much fun): day long meetings every week about how joystick movement feels, some sales coming constantly and arguing long about where company logo has to be, etc. You can imaging. Secret was that joystick was ready about the time hardware for main system unit was ready - iow it was finished already when all meetings took part. But since it was unusable w/o system unit - it was considered part of single project, rather than separate project. We were ready on dead-line. What we did one way or another highlighted that all those meetings were waste of time. "Nevermind" as my manager used to comment on that.
That has taught me that in some companies managers in fact has to do lots of work to allow mere developers to do the job. When you look inside - it might look like scam. But as long as job gets done and company has something to sell - that's okay. In big companies it is just too much interested parties - some-one has to keep them at bay, so that normal work can go on.
I wonder if that would hint Apple to licesning Windows Media for Mac OS from M$.
;-)
;-)
P.S. Thou, I think that's impossible, due to Apple's policies on QuickTime: ability to play files created with older version of QucikTime. M$ likes to drop stuff as they wish. Apple somewhat cares more about customers' experience.
P.P.S. I keep that WMP for Mac on my iBook - sometimes I'm getting WMV files from Wind0ze people. But, Thanks God, many now came back to DivX/XViD with MP3/AAC packaged as AVI or OGM. So most of the time it works now
P.P.P.S. I beleive, WMP for Mac always had the "Beta" moniker. Thou I think we can hardly call that "being released". Anyway, VLC is better. It would be good, if one day it will support WMV9 - but otherwise it is not something that I miss much. Anime fansub groups release DivX which plays with no problems by all players
P.P.P.P.S. "Our focus really is in delivering the best experience to Windows customers." Then M$ has to drop that crap called Windows Media and license and bundle DivX. That's my best experience on Wind0ze.
Why people in wealthy countries never think of moving into developing countries?
;-)
Thou I haven't tried that by myself
Freinds of mine - married couple of Uni teachers - were working in North Africa (Marocco & Alger) for about 10 years. When they came back I only hear them complaining how much they earn here but can buy literally nothing. Situation in Africa was different: they were earning little, but most of things costed next to nothing. In Africa they were top - here they are just average. (They came back because both wanted to have children in homeland.)
What just reminds me about how overvalued, overpriced developed countries became with time.
As many have already speculated, HDTV quality is high enough to be ripped from analogue signal and converted back to digital. ;-)
Anyway, DivX/XViD image size is normally scaled down, compared to DVD resolution - iow we are already used to lower-than-DVD quality. And it is Okay
Case 1: you buy CD, rip it and share mp3s on net.
Case 2: you connect to net and download some mp3s.
In first case you are criminal, since you knew that CD content is copyrighted, disregarded that and exercised the right you were not given by copyright law.
In second case you are innocent, since you have no clue about who is owning the content. And it is duty of copyright owner to explain that to you and ask to stop illegal distribution.
Modern P2P networks have merged both cases together - since everyone is uploader and downloader at the same time.
Thou again: to prove your wrongdoing in first case, the copyrighted work must be downloaded completely from you - but not someone else. If you have downloaded file from 100 hosts - none of the 100 people behind the ip addresses has uploaded complete work. RIAA not yet tried to crack on BT or eMule networks: before that we wouldn't know what U.S. legal system thinks about such kind of distribution.
Grand parent was trying to tell different point.
MySQL is fast but it has sacrificed many features for the performance.
Big DBs they might be slower - but over all they do provide more features and (very importantly) robust performance.
I have worked with MySQL, MS SQL & Sybase.
Sybase of Linux left the best impression overall. I haven't digged much Sybase on Windows - so I cannot tell. It just worked. Server crashed (or was just accidently unplugged) - it was getting up with no fuss. Scripting in the beggining was very confusing but few day of learning payed back much.
MS SQL Server pares on features but performance-wise needs a dedicated server - even if load is low server is barely usable for anything else. We tried to run IIS and MS SQL Server on single machine but later abandoned that idea. After Sybase complete control from command line, MS's love to silly configuration GUIs was bit irritating. (Especially when you still end-up firing regedit) Later on we discovered that MS SQL is very unhappy because of manual tuning applied - going back to defaults helped.
MySQL on Linux was pure fun. You can insert 1-2k rows in table per second using normal INSERT. (MS SQL/Sybase could both insert few hundred rows per second - for large operations it is recommended to use backup/restore facility) But then after say 100k records fun was beginning - MySQL was driving complete Linux box into eternal swap nirvana trying to index inserted data (which of course didn't fit the RAM). Who is guilty? Linux which over last 4 stable incarnations only now ca barely effectively handle files bigger than RAM? Or MySQL which wasn't expecting that someone will try to create table for 100k rows? I do not know. Then you did something like that "SELECT * FROM a WHERE a.a=1 or a.a == 3" over that 100k table - and you are going to drink some coffee - MySQL would need 10-20 minutes to finish the query. Why? - "OR" operation is not optimized and not a priority for developers. MySQL development was always very focused: if you can achieve something easier by hack elsewhere - do hack elsewhere. That's the response I got from Monty on my OR'd select problems: don't do selects with OR, instead do OR manually in your application. We later had to port complete stack from MySQL/Perl over to MS SQL/ColdFusion. It wasn't nice - I still have recurring nightmares of ColdFusuion4 scripts - but it worked Okay.
Well that was situation with MySQL back 5 years ago. For big systems we were using MS SQL/Windows & Sybase/Linux - for small stuff, especially for VERIO customers, MySQL was standard. You definitly do not need something big like Oracle for 3-5 tables barely exceeding 10k records - and that conf covers about 75% of private B2B portals.
Yeah... You seems to underestimate them.
What they want? I'll tell you what they want. Subscription service you cannot unsubscribe from with Britney Spears on every channel all the time. They do nothing and you pay them money. All the time. And you have no escape. Even if you do not watch - you still pay. That's what they want.
P2P file sharing and Internet gave people that "always on" feeling. But people refuse to pay fees **AA wants to charge them. **AA still lives in "on demand" times (of CD sales) as opposed to "always on."
Right now legal "always on", for example, is music TV channel, where you pay your subscription by watching lengthy ads. They have screwed radios and TV stations with lengty contract on how they can use the content. Now **AA want something like that for Internet. But Internet, created by Unix hackers, just lack proper ways for them to regulate how and what you are listening.
iTunes is sort'a compromise of the "always on" and "on demand" systems. But subscription services are what **AA truly commiting to: sustainable revenue stream. You know, artists are not stable, it's hard to do business with them.
Is it only me who is missing "Find Next"?
Seems like OOo people over-ab-used GNOME on their desktops...
When company found to be a monopoly, regulation body would be created. On the committee would be people from the company, its competitors and independent professionals from application field.
Such committee will oversee all actions of monopolist for some time (depends of the court order). Competitors can submit their complain and wishes - committee will decide on them what to do. Such committee can effectively over time put more restrictions on M$ - but only in line with court decision.
Deutsche Telekom (DT) of Germany is monopolist and regulated in that way. Court ordered them to open part of network for smaller ISPs thru number of connection points. Later on smaller ISPs have complained that DT has made the connection points practicly unusable to provide any good level of service to customers. In response, regulation body has investigated and confirmed that to be the case and has ordered DT to provide more connection points. After original court decision there were about 10 independent ISP here. Now - after regulation body intervention - about 50.
IOW, If M$ will submit incorrect or out-dated documentation, qualified people would notice that immediately. They will be involved in that process directly and report M$ incompliance.
How true. To the last word. Except that Linux use requires IQ130+ - my IQ is about of level of collective intellegence of slashdot population.
Seeing overtime how corrupt Europe is make me very sad. And creation of EU has just created many more cash cows the bureaucracy can milk - just recall money laundering scandal with EuroStat which 3 years later now is still nowhere to any resolution.
I think have M$ played the proper political tune, it could have delayed the case as long as it wanted. But M$ as true american company tried to muscle too much, what didn't frightened local politicians. Instead they have seen opportunities to make good careers themselves. M$ didn't get the difference between U.S. and Europe. It didn't expected that some will choose career later over money now.
As things now, M$ will have to throw some bone to the angry dog. Probably it will be considerate enough to make a friends with its owner.
P.S. On other side that might be a delaying game of M$. Delay long enough til when the APIs in question would become obsolete enough. Or even better til when Windows Server line will be dropped in favor of Vista Server/whatever but not "Windows Server". And they are off of the hook.
"""I just had an interesting thought that your comment about HTML+CSS prompted."""
"""What if the presentation is the content?"""
Well this is why XML's style sheets are XML - XSL.
And even more. Transformation of XML styles described in XML too. And they are standartized too - XSLT.
So you can transform style. And you can transform transformation. XML has enough rope for everyone.
And true American parents will be able to protect their children from the filth of The Real World: for example all scenes of violence will be automatically replaced by big red [CENSORED] banner right in the kids' brains!
Matrix! please have me!!!
I think RIAA/MPAA must build another country: with 100% secure content distribution, with 100% managed internet, with total protection of all those precious intellectual propery rights. And distribute precious content of theirs only there. Then they could run commercials of this land trying to persuade people to move in: "Joe Six Pack(TM) moved to WonderLand(TM) and patented letter A(TM)(P) - and now(RM)(pat. pend.) he is billionaire(TM)(P)! Now he has enough money to buy s/h TV set!!! WonderLand - land of new opportunities(RM)(TM)(C)(P)!!!!! [This ad(TM)(C)(P) in part brought to you by Joe Six Pack Eterprises(TM)(pat. pend.), exclusive(TM)(pat. pend.) licensor(TM)(P)(pat. pend.) of latin(C)(P)(RM)(TM)(pat. pend.) alphabet(TM) in WonderLand(TM)(pat. pend.)(RM)(pat. pend.)(TM)(RM)(pat. pend.). Reg-No: 8000-235435248-2343234. DRM-Id: 4334242-234234-2342342. Watermark-Id: 4343423456524234. MD5: 2323234542342424234234234. CRC32: 12123432. By watching this ad you give up all your rights in favor of Joe Six Pack Enterprises of WonderLand. FBI Warning: For illegal use of the ad, you might be sentenced for up to 50 years of life in WonderLand.]"
One was giving short insight into ODF v. M$' DOCX. (*)
1 44611543
ODF is flow based a-la HTML.
DOCX is record based a-la files generated by Write (.wri) & WinWord' text changes stream.
Application of styles is very different. Even if conversion of text can be made, conversion of styles is almost impossible. In DOCX styles are more or less inlined - ODF was redesigned by OASIS with styles to be more like HTML+CSS.
Basicly, M$'s concept boils down to "anything can occur anywhere in document". ODF hence standard is more strict.
Additionally M$ has special support for ActiveX: embedded objects will be stored as binary dump in middle of XML documents. (E.g. all pictures and files inserted from outside in M$ universe are ActiveX objects.) I'm not sure who ODF pares with embedded content, I can only hope OASIS - unlike M$ - have put XML to good use. After all, unlike M$, ODF includes vector and raster graphics too.
(*) http://www.groklaw.net/article.php?story=20051125
That's actually why most pirated DVDs and movies have that warning/copyright info deleted. (*)
I normally recognize pirated stuff from original by presence of the overlong legal statements. True pirates normally do not reproduce it.
So if you have bought pirated DVD, you might see no warnings whatsoever. And copyright statement could be (intentionally made) incorrect too. I'm not sure how court would regard distributing copies made off pirated copy. Technically (if you stopped distributing upon noitice from copyright holder) you have seen no copyright statement so you are not guilty. Guitly would be the guy who sold you the original pirated copy.
(*) This is where actually DRM-like crap started creeping into DVD standard: MPAA wanted a "cannot skip" flag for particular section of video. Ever tryed to skip that green/red warning? - most DVD players will err with "Not Permitted" message. (In the time, everyone was afraid that cheap retailer might use that for advertisments. Thanks God that didn't happened.)
I'm voting for VIM. (Just for record: I used vi once on Solaris - it's pain. So please please please all those Linux users do not mess vim with vi. If you have used only Linux that would mean you never really seen horrors of vi - all of OSs of GNU/Linux fame use vim instead of vi.)
I've seen people used to use emacs. But I've never seen anyone who really knows what is going on inside. Or truely understands how some things work. People just got used to Emacs - not understood it. Emacs is like religion - you cannot understand it by definition.
I like VIM. It integrates easily with Unix tools and shell.
In contrast, Emacs integrates with nothing - it has replacement for all the tool normally one would expect to be implemented on OS level.
Emacs is great tool, but learning it with all its modes and buffers could fill one's full-time position.
GNU - GNU's Not Unix, and Emacs is the confirmation. Try (re)configuring VIM and then try to configure Emacs. Emacs's init.el/friends always remind of sendmail.cf.
VIM doesn't have all the bells and whistles - but it does one thing and does it well. Emacs IMHO does many things - but does them averagely. After several years of unsuccecful tries to adopt almighty Emacs, I gave up on it. I know VIM and Bash - and can already do my job. With Emacs, regardless how much you already know Emacs is always ready to provide you with many new problems and obstacles. (Once I have tried to teach Emacs to use tab instead of spaces. It's very long and sad tale with no happy-end.)
VIM is simple and only hard thing to learn there is Insert v. Control mode. Once you got that, you can work already. I do not like it much - but I definitely like it more than cascaded N time nested M character long shortcuts of Emacs. VIM's simple shortcut structure also IMHO friendlier to touch-typists.
From programmer's POV, CC-mode is definitly superior to what VIM has to offer. But on other side, if something in CC mode doesn't fit your needs you might end-up killing two weeks to figuring out how to change that. VIM was hard to learn to use effectively. But once you learned it - due to its easy integration with OS - you find that you can make tasks you did before in shell better/faster/easier. (You can call Perl from VIM - or you can call VIM from Perl - choice is all yours.) It doesn't nullify your experience with Unix - like Emacs does - but rather magnifies it. IMHO.
I'm so tired of (ab)use of abbreviation 'IP' to denote something nonexistant.
There is only one true IP and it is STD5. Period. Discussion closed.
There is unfortunate side-effect.
Sony Classical is one of the biggest classical recording & CD publishing companies. If you will look into classical guitar (for example) you will find that about 90% of artists are published by Sony Classical.
I already hate SONY - with just this one little exception...
P.S. I recently bought a pile of CDs from Sony Classical, expecting to find a lot of (CP) markers, DRM toolkits, root-kits, thousand clause EULAs, etc. To my disappointment there were none included, even with SACDs.
I've been on one exhibition where some stats were drawn.
1st. Year 2003. In my application field, in North America (U.S. & Canada) ~650 start-up were founded. Europe? - 25!
2nd. My company - start-up - has made on exhibit to top 10 european start-ups. How? BY DEFAULT! We haven't yet managed to produce or sell something! It's just there were no other start-ups to compete against.
I'm living in Germany last 4 years - I yet to seen any progress at all. My poor home country Belarus - classified as "poor", "third-world" & "dictatorship" - sees more investments in development than "rich" Germany.
"Stagnation and protectionism" are two words I can use to describe local social and political systems.
And SAP actually is traditional German business working on traditional German principles. IOW. If complete idiot was hired - he will never be fired. (Competence of personel in Germany is really last thing anyone cares about. I'm working for third company and nobody - except staffers - ever looked into my CV.) All bugs are there to stay, since it can break numerous customer applications. 'Customer feedback' is something mythical, non-existent and ignored. Everything what have workaround is considered to be not bug, but feature. Ergonomics (it's over all German) doesn't ring any bells. Thick unpenetratable wall of management, secretaries, sales, service peresonel effectively shields any knowlegeable engineer from ever communicating with customers. And so on. I worked for similar company for some time.