Questions and Prospects of C Builderx
Raptor [Mental Studio]
Http://eental.mentsu.com
(One, problem article - on)
The original intention of writing this article originated in the QQ group of C Builder, built in the Yutong Book Brothers, on December 11, 2003, chatting with the book, the buddy, Ccrun, etc., a chat ( Chat records see "Discussion on C Builderx"). It is mainly to talk about the status quo of BCB / BCBX in the future, because I feel that it is helpful for friends who use BCB, so I will sort it with the support of the fox brother. The content of this article is as good as the topic, first talk about BCBX's current problems (including some less suitable practices in this product transformation), and then talk about it on its own point of view, although It is the prospect of BCBX. Of course, these are some of our airs, how will Borland do, it is not us.
More than two years ago I once envisioned that Borland will launch a highly integrated development tool. I called Borland Studio (see Hosted "Borland Studio?"), And later listened to Li Wei, Borland is developing code for Galileo. Tools will be a more powerful stuff than the Borland Studio I envision. However, very unfortunate, the so-called plan is not changed. Today, these stuffs have become bubbles, I don't know when Borland can give us such a surprise. (Supplement: Now it seems to have a hope, C # Builder is BDS1, Delphi8 is BDS2, this BDS is said to be Borland Development Studio, maybe BDS3 will give us a surprise)
As far as the current situation, Microsoft is already iron-hearted to transfer the Windows operating system to .NET, the future Windows operating system will no longer provide new Win32API, which will be available in .NET. The original Win32API is just reserved for compatibility. This is very similar to the case where the MS SQL 6.5 is migrated from MS SQL 6.5, and DBLIB in MS SQL 7.0 is only provided with a 6.5-compatible part, and the new features are provided in OLE-DB, and when MS SQL 2000, Dblib almost no longer used. MS SQL's transfer successfully achieved a leap and completely got rid of Sybase's shadow, so Microsoft should still play it in .NET.
In this case, the native development tool under the Windows platform is getting microphone, which is undoubtedly a major choice for Borland. So Borland has launched four main products this year (2003): C # Builder (Sidewinder), C Builderx (Tomahawk), JBuilderx (REVEILLE), and Delphi8 (Octane) just released on Borcon2003. From the surface, in addition to C # Builder is a new product that is introduced with Microsoft Visual Studio.Net 2003, the other seems to be the continuation of the original product, especially Delphi 8 extends the name of Delphi, and The X of JBuilderx launched after JBuilder9 is also very easy to understand as the "10" of the Roman number, then Is C Builder 是 C Builder 7? Why not call C Builder 7 or C Buildervii to call C Builderx? Will it be the legendary Borland Studio or even the prototype of Galileo? The greater the so-called hope, the greater the disappointment. When C Builder is still in the beta stage, many C Builder users can't wait to download Tomahawk to seek fast, I am the same, but the result is great. This has caused great repercussions in the BCB user group, which is more famous, such as the 9CBS BCB version of the Aweay brother's article "Borland listened to you", "
Although Borland China's left loudone has been awarded to BCBX, it is difficult to accept. BTW: According to the saying of the brother, Galileo has become the name of the unified IDE used by Borland's development tool (C # Builder and Delphi8) under the .NET platform, it seems to have a hope for Borland Studio.
It is undeniable that Borland is necessary in this transformation in C development tools from future development perspectives, and it is also the right practice. For details, please refer to the articles of the left brother ("answer about Borland C Builderx" or "Programmer" 2003 No. 12). But still point to Borland's error practices in this transformation and aspects of improvement. Those who we can't say representative Borland, but at least Borland's hardcore users, our opinion should be more representative, can say no exaggeration: If Borland can't make a corresponding improvement in the future, I am afraid that it is difficult to have a chance to turn over in the C field.
First, the way BCBX is launched is very unsuitable. The most important thing is to provide a smooth transition from BCB to BCBX. Borland has always marked its own feature: bringing new technologies to customers, creating new value, while protecting customers have investment. In the past, she has always been doing this, but this time is not done. This is a very serious mistake.
Historically said (see Li Wei's "Borland Legend" book), this error Borland has made many times, each of which causes a lot of loss in the user. Take C development tools as an example, the first time the OWL2 of BC4 is incompatible with the OWL1 of BC31, and the BC4 itself is poor quality, causing the first large-scale loss of Borland users. This incident eventually makes Borland from first-class software. The company has become a second-class software company. Although it has launched a quite good BC451, it is too late, returning to the heavens. Until BCB appearance, a part of the user was grabbed as the first RAD C tool, but for various reasons (one of course because the VCL used by BCB made it hovering under the shadow of Delphi), it has not been able to reproduce the BC31 The brilliant. It is impossible to imagine that Borland spends such a lesson of such a big price, but I still don't know how to absorb it, I have been less than ten years and make the same mistake again. When I first saw BCBX, I saw that I saw BC4 or BCB1 before a few years: I was fresh, big, then I was disappointed (I was not used in the middle), and then I started looking forward to the next The version can be solved (of course these two expectations, one is BC451- Unfortunately, the other is BCB3 - this is more fun, we are all looking forward to BCB2, but hope that BCB3, probably Borland jumped through BCB2 in order to match Delphi's version number.
In fact, Borland should avoid this dilemma and is not difficult. There are two ways: one is to temporarily embed a simplified version of a BCB IDE in BCBX, so that it can continue to support VCL's RAD development because of the C development tool, even at the beginning, a small number of cross-platform features (BCB) It is not possible to cross the platform), the problem is not big, because this part of the characteristics is originally to inherit the things under the Windows platform, not to mention the products that are highly required by JBuilder, which is not pure Java in earlier versions; Another method is to provide a BCB7 as a transition as Delphi. Because from a certain point of view, Delphi7 has improved to Delphi6, so some people play Delphi7 actually Delphi6.5. As far as the current situation, Delphi's transformation is undoubtedly more than the BCB, and Delphi8 provides .NET development and compatibility VCL development (vCl.net). From the situation of BCB, the only explanation is that Borland is not paying attention to BCB. In fact, in these two methods, individuals think that there is a better way later. Because although now .NET has a big way, it is certain that the native development under Windows will last for several years. In this time, BCB7 and BCBX are parallel (because in most cases, the conflict of the two is not very large) The Borland occupied C development tools is very favorable, and there is a BCB7 such a blank product (because most developments are going to .NET) can also bring a certain amount of income.
Although Borland as a company that cannot be too big, there is not enough resource to maintain a too long product line. If you want to join BCB in BCBX or develop BCBX and BCB7, it will be difficult. But as a new field of products (in the platform neutral C development tool, there is no product like BCBX, even if the legendary Eclipse CDT is temporarily unable to become a climate), slightly delayed some publication does not It will cause too much impact, but in turn launched a lot of mature products, it will cause serious back consequences. In this regard, Borland is also a lot of bitter, the most typical is Delphi4. And if BCB7 is used as a buffer, BCBX delay is published, and it will not have much impact on the market. But now Borland has been doing this. If you want to save, you only have an article on your product, so that Borland's own way you can't turn back - you must introduce a "available" BCBX as soon as possible. This "available" requires much higher than the BCB7 buffer.
(to be continued)