I've been using JSON (JavaScript Object Notation www.json.org) for a couple of years now for all my JavaScript application programming frameworks and it's paid off extremely well. I can easily express complex data structures in an extremely compact format that is easier to read than XML and is about 40% smaller. I use it for program configurations to hashmap type data structure definitions.
The beauty of JSON is that it is natively interpreted by both JavaScript and Python making it very fast to parse.
Sure JSON does not allow for complex meta definitions, but in most programming designs like XML but more often then not, XML is the equivalent of cracking open a walnut with a sledgehammer.
With proper coding you can use this meta definition model with any language.
True, Apple's iWork is designed for the masses but it does step on two thirds of Office-X. And I agree, OOO has features that the masses would not directly use but they could benefit from them without even being aware.
Where OOO wins over all other
The real important feature of OOO is it's programming framework though. I built several applications that ran solidly for years on Microsoft office 97 and VBA. These apps save millions of dollars in labour and mitigated losses for my company and with just coding in Excel and a bit of Access.
Unfortunate, but expected, most of these apps crashed when my company switched to Office 2003 which includes vb.net which is not backward compatible with older VBA code.
I have taken a very close look at the Object frameworks under OOO and they are not only unified (similar across all app components) but they are intuitively layed out. Anybody who can code in basic language can easily build an application that can do amazing things. And much more supportable and extensible than any Microsoft deployment. Another nice thing is that you can program in Python, JavaScript, Java, C++, TCL...
It is possible though that OOO does significant changes to their frameworks that will depricate or kill code like Microsoft has done so well with vb.net but it is my hope that they have learned from microsofts mistakes.
These features would not be used by the average joes out there but entry level hackers could easily make their start with OOO. VBA is where I made my start in hacking career as a hobbiest (not that I ever plan to write another piece of code on MS office ever again)
Really, it's not about Apple releasing their own version, I just hope that they can somehow shoot some money (back door like Microsoft/SCO way would work just fine)
Without a spreadsheet, I will not bother using this tool set. It is apparent that their marketing groups is focussing on a very specific group of individuals: Managers and marketers and their immediate reports. Any real grinding work is acomplished in a spreadsheet app.
Apple probably skipped the spreadsheet so they don't step on Microsofts feet with Office-X by leaving Excel the "real"* only choice for OSX. * - Besides using NeoOffice-J or OOO via X11:p
Note to Steve Jobs: Plesae expand your reseach in the area of native OOO support. Keeping Microsoft happy is keeping many Linux users (ex-microsoft users) from switching to OSX! If you support a native Open Office you will open the door for millions of switchers from Linux and Windows to OSX. And these users are very vocal - they will open the door for all their families and friends to switch over too. Just with my family alone, I have about 20 computes that I know of that I am telling the owners to wait for native OOO on OSX.
BTW - Probably best to leave out the marketing reps when you have such cost to benefit analysis done though! OOO is the future of office apps. PERIOD.
Unfortunately OSX does not render heavy swing based apps very well unless you have a decent video card to leverage Quartz Extreme. I found that jEdit on my older iBook (500mhz) as with other swing apps was too slow. If I had a faster mac, I'd use jEdit for sure. jEdit is probably better than BBEdit and Ultra-edit in many ways.
KDE's Kate and KDevelop ran very fast on all boxes that I've tried but you have to install the massive KDE framework to get these.
I have been keeping an eye on this project for a while and it's looking very nice. A cygwin port would be all that I need to keep me happy. I've been looking for a MDI based editor that will work on OSX and win32 for some time. But since my iBook died, I've been just scripting in Ultra-Edit.
1) Improved session cookie handling
-- https://bugzilla.mozilla.org/show_bug.cgi?id=11722 2
-- This is a major bug in the architecture of mozilla that needs to be fixed. Unfortunatly it is very encompasing and as such there are not many who are brave enough to tackle this one.
2) Message Box functionality
-- This would allow for functionality similar to but not exactly like IE's MsgBox.
-- Since alert, and confirm don't always cover all situations, it would be nice to enable yes/no, yes/no/cancel and some others.
-- Also fix the whitespace contraction problem with alerts (https://bugzilla.mozilla.org/show_bug.cgi?id=5034 8)
-- Allow for copying from the message (alert) dialog box. In IE you can hit ctrl-c and the content of the alert is copied into the clipboard (cool feature)
3) Access to print layout features like in IE
-- In IE, with the right evil Active-x control, a page designer can change the formatting of the printing via JavaScript. This is an awesome feature that we depend on for outputing professional documents from the web browser. However, Moz is not capable of this one. Maybe this could be done with an extention?
4) Better memory management in framework
-- In general, I find that Mozilla does not do a very good job of freeing up memory. Over time, when doing lots of refreshes, eventually Mozilla will take over 100mb even if only one window is left open. Better garbage collection is probably in order.
-- Right now, I have to shut down all Mozilla windows and re-open them from scratch to bounce the memory footprint.
5) Corporate distribution ability
-- A deployment of Mozilla that will enable a corporation to control and lock down which extentions and features are enabled. This feature is really for security reasons and I am sure that the Extentions engine is probably going to be the breaking point for may large groups to embrace Mozilla. If an IT lead can specify that the mozilla on the image will have xxx extentions and can routinely update these extentions or add/remove them, they can lock down this issue with mozilla. No security lead would be unhappy with this feature.
-- Maybe an extention service wrapper extention can be enabled for this one.
I spent a bit of time the last couple of weeks checking out the exibits at the Ontario Science Center on genetics which is probably one of the best exibits they have ever put together. In looking at the concepts of DNA/RNA/... A thought came to me - Why not build a programming language coding framework that is based on the strict constructs of genetics. The language/framework would implicitly have serializability of all structures and could allow for generation of truly extensible components. The basic concepts of highly structured data frameworks is growing (ie. http://nakedobjects.org) but why not pull these constructs one step away from the business data and bring it to the business logic or core application coding level.
Anybody know of such coding or at least theortical hacks out there?
Actually, Scarab (http://scarab.tigris.org) uses Maven under the hood. It also uses Velocity templates and many other really cool Apache hacks.
I checked out Trak and it looks nice. Cool that it integrates with Subversion. However, the issue tracking component is really very generic when compared to Scarab.
Although Scarab is missing a few key features (thanks to Collabnet hiding commits of these features to tigris.org) it is still very usable. The only thing I see with it is the customization workflows for defining custom issue meta constructs is a little poor (I'm sure that the full `collabnet` version has these fixed) but its nothing that a little fancy sql scriping can't work around.
If my team was large enough to use and enterprice level Collaborative SVCS, I'd choose collabnet any day, their web based interface is very well written, intuitive and fast. But like Rational, PVCS... it's very expensive.
I've used Scarab and subversion and both applications rock. Although scarab is quite alot more to set up, It's web based interface is way cleaner than bugzilla and easier for non-hackers to understand. The beauty of subversion is it's simplicity and it's ability to integrate with external systems via it's scriptability and `hooks` trigger framework.
BTW - I've been looking for an open source collaborative software development setup/framework/system that allows for integration of the version control system, the issue tracking system, document management... Anybody know of such a beast? Not the kind of thing one does on their spare time outside of their day job:]
I've seen what the tigris guy's are doing with scarab issue tracking system via velocity templates and I'm very impressed. The recursive development with velocity is much faster than JSP's and allows for very complete scripting.
I currently am maintaining and building onto a TCL based templating engine that is albeit limited in some ways but still rock solid and fully extensible.
If full JS support is available for your users, you could always dump the server side templating model and switch to a server dump to JS native data structures (arrays & json) and just do your templating in Java Script. I've got ton's of this code in production and it's way faster than doing server side templating and the deveopment cycle time is considerably less. But you really need to set your browser support to just Mozilla (1.3+), IE (5.5sp2+) and maybe Konquorer (safari) to do this kind of DHTML application programming well.
I've dumped JSP's about a year ago and now have a very simple servlet based template model.
This really means that every source off off the power grid needs to be isolated via a capacitor bank to remove any harmonic effects from bad electrical/electronic devices.
The cost of the power companies to put in such filters is not cost competetive. It's still probably cheaper to just run a fiber line beside the power line.
My brother actually built a simple version of this back in the mid 90's as an RCC project where he made a distributed stereo system that went through the power lines of the house. It worked unless you used our microwave, blender, dimmers...
I used copernic way back when their web based search tool still worked with google. I found that their tool greatly simplified my web searching workflows. Although, when they switched their UI to the XP look and feel, I lost interest. Besides loosing google groups searching, The user interface went too many steps backwards to be comfortable to use.
It's nice to see these guy's back on the map again.
When you port the 'core parts' out of J will you be moving to Obj-C or directly to C/C++? I would assume that an Obj-C/C/C++ port would be best completed while using Python as your OOP and proceduraly prototyping layer, Obj-C as your core framework layer and C/C++ as your performance/optimization layer.
BTW - How is the performance of Obj-C over Java, from what I have seen, Java is getting very close to the performance of C'ish languages but still with a large resource footprint, particuarily memory.
I'd be very interested in such hacks whey you decide to go that way.
I'd love to have one of those babies to sneak into my employers network to host my development hacks and scm repositories. Being somewhat familiar with bsd roots of OSX, I am currently hacking on FreeBSD but there are much more binary distribs on OSX for app's I'd like to work with.
I't would be great to have these boxes in other applications like at my family restaurant and bakery. I could build a very low cost Recipe, scheduling, POS... system and all running Obj-C, Python, Java, JavaScript.....:]
I hate suggesting things that support continued use of IE but since we are talking charity here it is:
You can probably wrap the browser session with a frame navigator (like ask jeeves...) where the controlling frame has all the navigation buttons and necessary menu items and even an address bar. When the browser starts up, hide all top menus and only show the buttons and menus you want them to see via DHTML. You could even create a bookmark based system using DHTML and some simple server side storage. The only difficulty is that you would need to put an authentication layer to resolve the current user although there may be a way to resolve this with an active-x plugin or even native.
Although you are looking at a bit of coding here, I know that you could use a citrix frame to navigate any IE based site in this way.
If your deployment requires more than about 30-40 concurrent users, depending on complexity of app,an Win.net deployment would probably require multiple computer arrays with complex gateway framework. M$ is not known for building such large frameworks well. Instead they focus on making smaller deployments fast to write in and fast to run.
On the other hand, with Java you can deploy on a nice juicy 10 processor *nix beast with terrabytes of stripped raid storage. By going J2EE on these systems you can get very high concurrency on one machine. You must have multiple boxes to acomplish this with win32. More computers == more complexity == more problems. With J2EE and Java RMI's (old but good) you can get very good multi-processor utilization.
Also, J2EE is better at deploying across multiple server farms such a linux array. By using multiple rmi's (one/linux box/processor) and an RMI gateway, you can get a very powerfull, stable and secure system. I get shivers thinking of how Microsoft is able to do this.
From my experience, I am seeing more and more.net deployments that are having very serious performance problems even though they through tons of bucks at the hardware, they would all have been better off with J2EE. The performance is often related to both high and low concurrency.
If your deployments are less than say 40 concurrent users, win.net is a possible solution but will be less stable and secure than J2EE on *nix. Above this 20 concurrent users, hands down I would go J2EE. At this level, the cost of extra processors & memory is relatively low cost in comparison to the application development & deployment costs.
Spend a bit more money on the hardware and you will sleep well having J2EE cranking away.
When accessing subversion using web-dav, you must run Apache 2 as documented here [big page]
Using web-dav and subversion is a must if you need firewall transparency. Although I found that the http sniffers on my corporate network actually slowed down my repo so I use svnserve instead (no apache / web-dav needed)
Slashdot Mirror
The only things wrong with iBook over this is the cool flip screen and the size.
Although not as conveniently portable (yet) you can get an iMac Mini with a 12" lcd, key and mouse for about the same price.
When the price comes down, maybe it will be worth while taking a serious look.
Hmmmm.....
No Judge in his right mind would ever allow this blatant waste of legal resources to take place.
I've been using JSON (JavaScript Object Notation www.json.org) for a couple of years now for all my JavaScript application programming frameworks and it's paid off extremely well. I can easily express complex data structures in an extremely compact format that is easier to read than XML and is about 40% smaller. I use it for program configurations to hashmap type data structure definitions.
The beauty of JSON is that it is natively interpreted by both JavaScript and Python making it very fast to parse.
Sure JSON does not allow for complex meta definitions, but in most programming designs like XML but more often then not, XML is the equivalent of cracking open a walnut with a sledgehammer.
With proper coding you can use this meta definition model with any language.
json={
coolness: "high",
ease: "high",
nativeSupport:[
{type: "language", name: "JavaScript"}, {type: "lang", name: "Python"},
],
nonNativeSupport: "You name it"
}
JsD
That's pretty damn close to apple's finder window. Can they do that without issues from Apple?
Not that I don't like it. Looking at this, I may consider going Linux someday...
True, Apple's iWork is designed for the masses but it does step on two thirds of Office-X. And I agree, OOO has features that the masses would not directly use but they could benefit from them without even being aware.
Where OOO wins over all other
The real important feature of OOO is it's programming framework though. I built several applications that ran solidly for years on Microsoft office 97 and VBA. These apps save millions of dollars in labour and mitigated losses for my company and with just coding in Excel and a bit of Access.
Unfortunate, but expected, most of these apps crashed when my company switched to Office 2003 which includes vb.net which is not backward compatible with older VBA code.
I have taken a very close look at the Object frameworks under OOO and they are not only unified (similar across all app components) but they are intuitively layed out. Anybody who can code in basic language can easily build an application that can do amazing things. And much more supportable and extensible than any Microsoft deployment. Another nice thing is that you can program in Python, JavaScript, Java, C++, TCL...
It is possible though that OOO does significant changes to their frameworks that will depricate or kill code like Microsoft has done so well with vb.net but it is my hope that they have learned from microsofts mistakes.
These features would not be used by the average joes out there but entry level hackers could easily make their start with OOO. VBA is where I made my start in hacking career as a hobbiest (not that I ever plan to write another piece of code on MS office ever again)
Really, it's not about Apple releasing their own version, I just hope that they can somehow shoot some money (back door like Microsoft/SCO way would work just fine)
Without a spreadsheet, I will not bother using this tool set. It is apparent that their marketing groups is focussing on a very specific group of individuals: Managers and marketers and their immediate reports. Any real grinding work is acomplished in a spreadsheet app.
:p
Apple probably skipped the spreadsheet so they don't step on Microsofts feet with Office-X by leaving Excel the "real"* only choice for OSX.
* - Besides using NeoOffice-J or OOO via X11
Note to Steve Jobs: Plesae expand your reseach in the area of native OOO support. Keeping Microsoft happy is keeping many Linux users (ex-microsoft users) from switching to OSX! If you support a native Open Office you will open the door for millions of switchers from Linux and Windows to OSX. And these users are very vocal - they will open the door for all their families and friends to switch over too. Just with my family alone, I have about 20 computes that I know of that I am telling the owners to wait for native OOO on OSX.
BTW - Probably best to leave out the marketing reps when you have such cost to benefit analysis done though! OOO is the future of office apps. PERIOD.
JsD
[Use FireFox or Die!]
Unfortunately OSX does not render heavy swing based apps very well unless you have a decent video card to leverage Quartz Extreme. I found that jEdit on my older iBook (500mhz) as with other swing apps was too slow. If I had a faster mac, I'd use jEdit for sure. jEdit is probably better than BBEdit and Ultra-edit in many ways.
KDE's Kate and KDevelop ran very fast on all boxes that I've tried but you have to install the massive KDE framework to get these.
I have been keeping an eye on this project for a while and it's looking very nice. A cygwin port would be all that I need to keep me happy. I've been looking for a MDI based editor that will work on OSX and win32 for some time. But since my iBook died, I've been just scripting in Ultra-Edit.
From the screenies, they have come a long way.
Great work!
Mozilla unfortunatly does not support all the print layout CSS functionality.
e r_support/printing.html
See: http://www.westciv.com/style_master/academy/brows
I guess I would re-write my request for Mozilla then to support this css functionality.
Thanks for the info though.
No need for ActiveX, this is already possible using standards-compliant CSS/DOM+JS.
Where, how? I have looked before and never found such code or abilities in Mozilla. Please share :]
1) Improved session cookie handling2 2
4 8)
-- https://bugzilla.mozilla.org/show_bug.cgi?id=1172
-- This is a major bug in the architecture of mozilla that needs to be fixed. Unfortunatly it is very encompasing and as such there are not many who are brave enough to tackle this one.
2) Message Box functionality
-- This would allow for functionality similar to but not exactly like IE's MsgBox.
-- Since alert, and confirm don't always cover all situations, it would be nice to enable yes/no, yes/no/cancel and some others.
-- Also fix the whitespace contraction problem with alerts (https://bugzilla.mozilla.org/show_bug.cgi?id=503
-- Allow for copying from the message (alert) dialog box. In IE you can hit ctrl-c and the content of the alert is copied into the clipboard (cool feature)
3) Access to print layout features like in IE
-- In IE, with the right evil Active-x control, a page designer can change the formatting of the printing via JavaScript. This is an awesome feature that we depend on for outputing professional documents from the web browser. However, Moz is not capable of this one. Maybe this could be done with an extention?
4) Better memory management in framework
-- In general, I find that Mozilla does not do a very good job of freeing up memory. Over time, when doing lots of refreshes, eventually Mozilla will take over 100mb even if only one window is left open. Better garbage collection is probably in order.
-- Right now, I have to shut down all Mozilla windows and re-open them from scratch to bounce the memory footprint.
5) Corporate distribution ability
-- A deployment of Mozilla that will enable a corporation to control and lock down which extentions and features are enabled. This feature is really for security reasons and I am sure that the Extentions engine is probably going to be the breaking point for may large groups to embrace Mozilla. If an IT lead can specify that the mozilla on the image will have xxx extentions and can routinely update these extentions or add/remove them, they can lock down this issue with mozilla. No security lead would be unhappy with this feature.
-- Maybe an extention service wrapper extention can be enabled for this one.
JsD
[Use Firefox or Die!]
I spent a bit of time the last couple of weeks checking out the exibits at the Ontario Science Center on genetics which is probably one of the best exibits they have ever put together. In looking at the concepts of DNA/RNA/... A thought came to me - Why not build a programming language coding framework that is based on the strict constructs of genetics. The language/framework would implicitly have serializability of all structures and could allow for generation of truly extensible components. The basic concepts of highly structured data frameworks is growing (ie. http://nakedobjects.org) but why not pull these constructs one step away from the business data and bring it to the business logic or core application coding level.
Anybody know of such coding or at least theortical hacks out there?
JsD
Actually, Scarab (http://scarab.tigris.org) uses Maven under the hood. It also uses Velocity templates and many other really cool Apache hacks.
I checked out Trak and it looks nice. Cool that it integrates with Subversion. However, the issue tracking component is really very generic when compared to Scarab.
Although Scarab is missing a few key features (thanks to Collabnet hiding commits of these features to tigris.org) it is still very usable. The only thing I see with it is the customization workflows for defining custom issue meta constructs is a little poor (I'm sure that the full `collabnet` version has these fixed) but its nothing that a little fancy sql scriping can't work around.
If my team was large enough to use and enterprice level Collaborative SVCS, I'd choose collabnet any day, their web based interface is very well written, intuitive and fast. But like Rational, PVCS... it's very expensive.
:]
I've used Scarab and subversion and both applications rock. Although scarab is quite alot more to set up, It's web based interface is way cleaner than bugzilla and easier for non-hackers to understand. The beauty of subversion is it's simplicity and it's ability to integrate with external systems via it's scriptability and `hooks` trigger framework.
BTW - I've been looking for an open source collaborative software development setup/framework/system that allows for integration of the version control system, the issue tracking system, document management... Anybody know of such a beast? Not the kind of thing one does on their spare time outside of their day job
JsD
I've seen what the tigris guy's are doing with scarab issue tracking system via velocity templates and I'm very impressed. The recursive development with velocity is much faster than JSP's and allows for very complete scripting.
I currently am maintaining and building onto a TCL based templating engine that is albeit limited in some ways but still rock solid and fully extensible.
If full JS support is available for your users, you could always dump the server side templating model and switch to a server dump to JS native data structures (arrays & json) and just do your templating in Java Script. I've got ton's of this code in production and it's way faster than doing server side templating and the deveopment cycle time is considerably less. But you really need to set your browser support to just Mozilla (1.3+), IE (5.5sp2+) and maybe Konquorer (safari) to do this kind of DHTML application programming well.
I've dumped JSP's about a year ago and now have a very simple servlet based template model.
JsD
Has anybody ever seen an OSX box crap out on Steve? I have not heard of this or seen it.
Hmmmm.
Good excuses are still just good excuses.
JsD
This really means that every source off off the power grid needs to be isolated via a capacitor bank to remove any harmonic effects from bad electrical/electronic devices.
The cost of the power companies to put in such filters is not cost competetive. It's still probably cheaper to just run a fiber line beside the power line.
My brother actually built a simple version of this back in the mid 90's as an RCC project where he made a distributed stereo system that went through the power lines of the house. It worked unless you used our microwave, blender, dimmers...
JsD
I used copernic way back when their web based search tool still worked with google. I found that their tool greatly simplified my web searching workflows. Although, when they switched their UI to the XP look and feel, I lost interest. Besides loosing google groups searching, The user interface went too many steps backwards to be comfortable to use.
It's nice to see these guy's back on the map again.
When you port the 'core parts' out of J will you be moving to Obj-C or directly to C/C++? I would assume that an Obj-C/C/C++ port would be best completed while using Python as your OOP and proceduraly prototyping layer, Obj-C as your core framework layer and C/C++ as your performance/optimization layer.
BTW - How is the performance of Obj-C over Java, from what I have seen, Java is getting very close to the performance of C'ish languages but still with a large resource footprint, particuarily memory.
I'd be very interested in such hacks whey you decide to go that way.
JsD
I'd love to have one of those babies to sneak into my employers network to host my development hacks and scm repositories. Being somewhat familiar with bsd roots of OSX, I am currently hacking on FreeBSD but there are much more binary distribs on OSX for app's I'd like to work with.
..... :]
I't would be great to have these boxes in other applications like at my family restaurant and bakery. I could build a very low cost Recipe, scheduling, POS... system and all running Obj-C, Python, Java, JavaScript
Let it be, let it be...
JsD
I hate suggesting things that support continued use of IE but since we are talking charity here it is:
You can probably wrap the browser session with a frame navigator (like ask jeeves...) where the controlling frame has all the navigation buttons and necessary menu items and even an address bar. When the browser starts up, hide all top menus and only show the buttons and menus you want them to see via DHTML. You could even create a bookmark based system using DHTML and some simple server side storage. The only difficulty is that you would need to put an authentication layer to resolve the current user although there may be a way to resolve this with an active-x plugin or even native.
Although you are looking at a bit of coding here, I know that you could use a citrix frame to navigate any IE based site in this way.
Good luck.
JsD
If your deployment requires more than about 30-40 concurrent users, depending on complexity of app,an Win.net deployment would probably require multiple computer arrays with complex gateway framework. M$ is not known for building such large frameworks well. Instead they focus on making smaller deployments fast to write in and fast to run.
.net deployments that are having very serious performance problems even though they through tons of bucks at the hardware, they would all have been better off with J2EE. The performance is often related to both high and low concurrency.
On the other hand, with Java you can deploy on a nice juicy 10 processor *nix beast with terrabytes of stripped raid storage. By going J2EE on these systems you can get very high concurrency on one machine. You must have multiple boxes to acomplish this with win32. More computers == more complexity == more problems. With J2EE and Java RMI's (old but good) you can get very good multi-processor utilization.
Also, J2EE is better at deploying across multiple server farms such a linux array. By using multiple rmi's (one/linux box/processor) and an RMI gateway, you can get a very powerfull, stable and secure system. I get shivers thinking of how Microsoft is able to do this.
From my experience, I am seeing more and more
If your deployments are less than say 40 concurrent users, win.net is a possible solution but will be less stable and secure than J2EE on *nix. Above this 20 concurrent users, hands down I would go J2EE. At this level, the cost of extra processors & memory is relatively low cost in comparison to the application development & deployment costs.
Spend a bit more money on the hardware and you will sleep well having J2EE cranking away.
JsD
Dooooh!
.... :p
You are so right.
I feel so ashamed
I had to disable my M$ firewall after installing XPsp2 because of daily crashes. Since disabling, I no longer get crashes.
Use another professional FW product.
JsD
[dreaming of not working through windoze on my corporate heat generator]
When accessing subversion using web-dav, you must run Apache 2 as documented here [big page]
Using web-dav and subversion is a must if you need firewall transparency. Although I found that the http sniffers on my corporate network actually slowed down my repo so I use svnserve instead (no apache / web-dav needed)