The most lovely Delphi message posted.
I had the most weirdest dream...
What would DelphiHater do, if he was in charge of Delphi and wanted to make Delphi successful?
First thing, DelphiHater would let many old timers leave Embarcadero. The kind of bad behavior on the newsgroups and antics would make many new customers uncomfortable with them.
Second, get rid of the blowhard Delphi fans. Sure, we like them, but when you call newcomers "idiots", "morons", and other expletives on the newsgroups, they have to go, or get warnings. So what if they are idiots? At least they asked for help. Get rid of all the troll-crossings so the new comers feel at home and ready to ask questions. At least that was the idea of Joel Spoolsky's StackOverflow.com
Third, cut-off the NNTP supply to google and then host it in-house, so that all the Adsense money can come into Embcardero, not some Usenet site. DelphiHater might consider suing some of the upstream providers that suck on Embcardero's Usenet for Adsense revenues and copyright (since they also provide extra services, like free downloads of copyrighted materials)
Forth, hire some more Quality Assurance and documentation. Why not ask some whitehouse interns to do Quality Assurance and Documentation? They seem to make sure the President gets quality service and kept upto-date on state of affairs.
Fifth, give money to fix bugs. If US$20 was given to fix all the damn bugs in Delphi, it would cost US$2,000,000 to fix (approx 100,000 bugs) Delphi, which would be cheaper than the US$100,000,000 a year Borland used to lose on Delphi. At least it would keep Andy J (of the Delphi Speedup Fame) and other hackers interested in Delphi.
Sixth, DelphiHater would consider buying over GExperts or CNPack, or even Castalia and free it from the GPL license and then build these features into the IDE.
Seventh, DelphiHater would make peace with Mark Miller of CodeRush and give him what he wants rather than some half-dead Editor Integration. BTW, get rid of that customized TXControl nonsense editor. There are plenty of Syntax-based Editor that Embcardero can use that is extensive, integrated and works with Delphi.
Eight, DelphiHater would hire fresh blood for the DCG (Delphi Code Generator) back-end. If the people from Borlando and CodeGearo cannot understand I64, then if new people can, get them to make a fresh decent DCC64.EXE that actually works. Maybe hire some people from QUT (Queensland University of Technology), the people who made that Pascal-clone in NET. I'm sure there are plenty of students who would be thrilled to make exciting career.
Nine, DelphiHater would tell Indy Pit Crew and team to stop this nonsense on newsgroups and get paid people to fix all those annoying bugs in Indy. Either way, either you give decent working Indy code, or just drop the whole Indy and get something decent, like beefing up ICS or RemObjects Soap or some TCP/IP library that actually works 100% during run-time, instead of crashing sometimes or working half-the-time.
Ten, DelphiHater would call to tender for another VCL Library for Internet. or even Scripted Delphi that works with IIS. Euro 600 per person for IntraWeb is waay too expensive.
Eleven, Make that spirit of Delphi award give money. A little money goes a long way to oil the wheels of Delphi...
Twelve, ask what happened to Team Spirit Nevrona. Their Rave report-writer is starting to smell. Either they start to maintain it, or FastReports included with Delphi sounds nice.
Thirteen, the Database drivers. Either open-source code, or give paid access (hint: Payment + NDA) to developers interested in the source codes for the Database Drivers or even BDE. Maybe someone will come up with something better than the Bullshit Database Engine (BDE) or even fix all those damn bugs in it.
Fourteen, pay people to answer question. Redeem points for prizes or free copies of Delphi. Then those lonely posts on the newsgroups will feel some love.
Fifteen, start to question suppliers who supplied the third-party found in Delphi, to make sure they are accountable for their work. It's no use giving out a Partner CD with nearly/almost useless products. BTW. Did someone do QA on the Partner CD?
Sixteen, instead of subscriptions, consider alternative licensing methods. Since the Chinese and Iranians use unpaid copies of Delphi, here's an interesting model. License by usage (checked by on-line activation). For example, pay US$2,000 and get a US$1,000 refund for license compliance.
Seventeen, there are already license managers that can detect Virtual Machine and usage. Since those license managers can activate and deactivate inside VMs, then why don't Delphi's license manager do same?
Eighteen, the Borland Certification is almost useless. It confers no benefits, no additional statuses and people who go for it, get no benefits. Either scrap it, or go for community model where people pass competency tests, or make the course-ware very public so more people get interested to learn Delphi.
Nineteen, Java has J2EE server. Delphi should have the same. Anyone for Delphi version of TomCat that serves Pascal script-lets? or Pascal scripts that bind to SOAP? Maybe Delphi JMI or Delphi ODBC adapters.
Twenty, ensure interoperability with basic standards, like SOAP, JMI, JSON and basic working XML parser that actually works. At least, upgrade Jeff Rafter's / Defined XML wrappers to work in Delphi 2009, get StreaSec II to work in Delphi 2009. or at least get unicode encryption to work, or so it produces same results as in NET and Java.
Twenty-one, put some power into the TurboPower-failure. There is plenty of code that still uses legacy TurboPower things.
Twenty-two, reward customer support staff based on feedback. Too many bad support calls, fire them. Maybe they are actually paid to do disservice?
Twenty-three. Buy over TMS or LMD, at least there will be 300 new components. Why not buy DevExpress and FastReports. There will be excellent grid, working report writer and decent edit controls.
Twenty-four. Buy over a working Profiler or Memory manager that actually works. Right now you cannot exactly profile without paying loads of cash, and you cannot check OLE string leaks, errors, or resource leaks...
Twenty-five. Fix the dumb Borland C++ Builder so that with more C++ compliance similar to Gnu C++ or Microsoft C++, people would say that's stupid, but read Andy J's blog how he hates C++ Builder. Then at least DelphiHater can use more C++ libraries than buy wrappers from Yunqua (The Delphi Inspiration).
Twenty-six. Where is BOLD? Why is it not available since Delphi 2007 Arch. had it and disappeared in Delphi 2009 Unicode version? When would a low-cost version be available for this?
The Delphi Ultimatum is really about business improvements.
Is Embcardero listening?
I had the most weirdest dream...
What would DelphiHater do, if he was in charge of Delphi and wanted to make Delphi successful?
First thing, DelphiHater would let many old timers leave Embarcadero. The kind of bad behavior on the newsgroups and antics would make many new customers uncomfortable with them.
Second, get rid of the blowhard Delphi fans. Sure, we like them, but when you call newcomers "idiots", "morons", and other expletives on the newsgroups, they have to go, or get warnings. So what if they are idiots? At least they asked for help. Get rid of all the troll-crossings so the new comers feel at home and ready to ask questions. At least that was the idea of Joel Spoolsky's StackOverflow.com
Third, cut-off the NNTP supply to google and then host it in-house, so that all the Adsense money can come into Embcardero, not some Usenet site. DelphiHater might consider suing some of the upstream providers that suck on Embcardero's Usenet for Adsense revenues and copyright (since they also provide extra services, like free downloads of copyrighted materials)
Forth, hire some more Quality Assurance and documentation. Why not ask some whitehouse interns to do Quality Assurance and Documentation? They seem to make sure the President gets quality service and kept upto-date on state of affairs.
Fifth, give money to fix bugs. If US$20 was given to fix all the damn bugs in Delphi, it would cost US$2,000,000 to fix (approx 100,000 bugs) Delphi, which would be cheaper than the US$100,000,000 a year Borland used to lose on Delphi. At least it would keep Andy J (of the Delphi Speedup Fame) and other hackers interested in Delphi.
Sixth, DelphiHater would consider buying over GExperts or CNPack, or even Castalia and free it from the GPL license and then build these features into the IDE.
Seventh, DelphiHater would make peace with Mark Miller of CodeRush and give him what he wants rather than some half-dead Editor Integration. BTW, get rid of that customized TXControl nonsense editor. There are plenty of Syntax-based Editor that Embcardero can use that is extensive, integrated and works with Delphi.
Eight, DelphiHater would hire fresh blood for the DCG (Delphi Code Generator) back-end. If the people from Borlando and CodeGearo cannot understand I64, then if new people can, get them to make a fresh decent DCC64.EXE that actually works. Maybe hire some people from QUT (Queensland University of Technology), the people who made that Pascal-clone in NET. I'm sure there are plenty of students who would be thrilled to make exciting career.
Nine, DelphiHater would tell Indy Pit Crew and team to stop this nonsense on newsgroups and get paid people to fix all those annoying bugs in Indy. Either way, either you give decent working Indy code, or just drop the whole Indy and get something decent, like beefing up ICS or RemObjects Soap or some TCP/IP library that actually works 100% during run-time, instead of crashing sometimes or working half-the-time.
Ten, DelphiHater would call to tender for another VCL Library for Internet. or even Scripted Delphi that works with IIS. Euro 600 per person for IntraWeb is waay too expensive.
Eleven, Make that spirit of Delphi award give money. A little money goes a long way to oil the wheels of Delphi...
Twelve, ask what happened to Team Spirit Nevrona. Their Rave report-writer is starting to smell. Either they start to maintain it, or FastReports included with Delphi sounds nice.
Thirteen, the Database drivers. Either open-source code, or give paid access (hint: Payment + NDA) to developers interested in the source codes for the Database Drivers or even BDE. Maybe someone will come up with something better than the Bullshit Database Engine (BDE) or even fix all those damn bugs in it.
Fourteen, pay people to answer question. Redeem points for prizes or free copies of Delphi. Then those lonely posts on the newsgroups will feel some love.
Fifteen, start to question suppliers who supplied the third-party found in Delphi, to make sure they are accountable for their work. It's no use giving out a Partner CD with nearly/almost useless products. BTW. Did someone do QA on the Partner CD?
Sixteen, instead of subscriptions, consider alternative licensing methods. Since the Chinese and Iranians use unpaid copies of Delphi, here's an interesting model. License by usage (checked by on-line activation). For example, pay US$2,000 and get a US$1,000 refund for license compliance.
Seventeen, there are already license managers that can detect Virtual Machine and usage. Since those license managers can activate and deactivate inside VMs, then why don't Delphi's license manager do same?
Eighteen, the Borland Certification is almost useless. It confers no benefits, no additional statuses and people who go for it, get no benefits. Either scrap it, or go for community model where people pass competency tests, or make the course-ware very public so more people get interested to learn Delphi.
Nineteen, Java has J2EE server. Delphi should have the same. Anyone for Delphi version of TomCat that serves Pascal script-lets? or Pascal scripts that bind to SOAP? Maybe Delphi JMI or Delphi ODBC adapters.
Twenty, ensure interoperability with basic standards, like SOAP, JMI, JSON and basic working XML parser that actually works. At least, upgrade Jeff Rafter's / Defined XML wrappers to work in Delphi 2009, get StreaSec II to work in Delphi 2009. or at least get unicode encryption to work, or so it produces same results as in NET and Java.
Twenty-one, put some power into the TurboPower-failure. There is plenty of code that still uses legacy TurboPower things.
Twenty-two, reward customer support staff based on feedback. Too many bad support calls, fire them. Maybe they are actually paid to do disservice?
Twenty-three. Buy over TMS or LMD, at least there will be 300 new components. Why not buy DevExpress and FastReports. There will be excellent grid, working report writer and decent edit controls.
Twenty-four. Buy over a working Profiler or Memory manager that actually works. Right now you cannot exactly profile without paying loads of cash, and you cannot check OLE string leaks, errors, or resource leaks...
Twenty-five. Fix the dumb Borland C++ Builder so that with more C++ compliance similar to Gnu C++ or Microsoft C++, people would say that's stupid, but read Andy J's blog how he hates C++ Builder. Then at least DelphiHater can use more C++ libraries than buy wrappers from Yunqua (The Delphi Inspiration).
Twenty-six. Where is BOLD? Why is it not available since Delphi 2007 Arch. had it and disappeared in Delphi 2009 Unicode version? When would a low-cost version be available for this?
The Delphi Ultimatum is really about business improvements.
Is Embcardero listening?
7 comments:
I really really like your post, home Rentcadero will listen and do something, but I doubt it, after all it's all about a few investors which wants a product that brings money, not an awesome product(which Delphi is) which brings money...
Keep up the good work!
You're welcome Dorin.
Time permitting I'll write some technical articles.
such as:
- DoJo Json + Delphi
- How do make a Delphi add-in to Visual Basic 6.
- How to understand Andy's J. patch to make your own patches.
- Programming Services with Citrix.
Good article!
Russian translation is here: http://keeper89.blogspot.com/2009/12/delphi.html
That's a weirdly constuctive post you have there, Mr Hater. Some observations about your points:
1. I can think of only four people in total who fit the category of 'old timer' and 'posts on the forums now and again'. (I'll define 'old timer' as someone who was a Borland employee at least since the time of D2006's release.) Given three never get into arguments on the internet, that leaves... well, he was particularly bad when he posted more regularly on the old newsgroups, indeed.
2. That'll be KW you're referring to I take it...? He is actually keen to help if you avoid non-tech, bizarrely enough.
3. Not sure what you're referring to - I thought the move to the 'new' forums broke this?
4. Well, they did seem to do this, but how they managed to make the D2010 help go backwards even still I don't know.
5. Maybe not - the main costs involved would be in verifying proposed fixes, not doling out reward money.
6. Possbly, *assuming* this won't affect IDE stability (I've been wary of GExperts in the past because of this).
7. I'd assume Mark Miller is long long - 'making peace' may have had meaning in 2004, but not now.
8. IMO, Barry Kelly coming on board was the best thing to happen to Delphi for a long time - anonymous methods, the extended RTTI and attributes seem largely down to him. Given Eli Boiling's blog posts about details of the Mac compiler development as well, I seriously doubt a lack of inhouse expertese is holding back the completion and release of a 64 bit compiler - rather, it's the conscious decision of management.
9. Leaving aside the personal attack (its not big and its not clever to do this behind a cloak of anonymity for yourself), I agree.
10. Maybe *any* Delphi solution here just isn't the best tool for the job...
11. Evidence?
12. Quite probably.
13. The BDE?! That said, IIRC, quite a lot of work was done on the DBExpress drivers for D2010.
14. Paying people small amounts is your Big Idea I see...
15. Maybe (I just ignore it myself).
16. If the existing registration method is hacked so easily, why won't your alternative be too?
17. Maybe.
18. 'Borland'?
19. No comment.
20. Wasn't a lot of JSON support stuff added to D2010?
21. Maybe.
22. Never used, so no comment.
23. No!!! That would kill off the third party component market straight off.
24. Sounds sensible.
25. Is there any real potential profit for Embarcadero in doing this though? For sure, it would be nice to have from a Delphi POV (though hardly necessary - if you want to use a C++ library from Delphi, just download VC++ Express and create a DLL wrapper). However, the legacy of the C++ compiler's underdevelopment when Borland lost faith in the product after BCB5 would make realising it expensive I assume, given backwards compatibilty and continued compatibility with Delphi and the VCL is a must.
Chris,
the personal attack to C was removed.
time permitting, i will answer your reply in detail
:)
i would like to add one point - Hire some people to document, clean up and fix some bugs in Jedi (JCL and JVCL). The Jedi library has so much potential, but it's unorganized, undocumented and has bugs. After the quality is improved it could be official part of Delphi.
Then again they can't even document Delphi itself, so this point is probably a pipe dream...
Post a Comment