Also, Jim, from a 25 year veteran in software development, myself: You have the right idea.
A rewrite is probably a formidable task, since this is an enterprise-level web app. My offhand guestimate is that it would take 2 senior developers, 1 fulltime tester, and one SOLID architect to pull off a full Free Republic implementation in 9 months to a year. Let us say, it is a year. Senior developers go about 100k, testers 60k, and an architect of the caliber you'd want is about 200k. That's 450k, and we haven't even talked about licensing development software and the equipment for them to develop on, much less the server farm itself. Figure 600k-750k to do a redevelopment.
John's system is, frankly, amazing for the price it's cost. it's obvious it's a labor of love. Plus, for him to have accomplished it in a bass-ackwards language like Perl, is truly mindboggling. (Poke poke, John.... lol.... told ya .NET was the way to go......lol.....)
Anyways, the plan you have adopted -- Keep up with the existing software, for it does work, and when needed, troubleshoot -- is viable. Objectively speaking, it is a decent plan.
There may come a time when the Perl system you have simply won't cut it any more. When that happens, I suggest a conversion to a robust enterprise solution. I favor C#.NET with SQL back end. It's battle-tested and well-regarded for enterprise apps. John would be a pro in it in about a month.
Oh my, I did not even consider this site was still using PERL for such a high volume website.
One plan for a future change is to keep the old posts in the depreciated PERL version as an archive and just start fresh with a new system/database with username/password imported from the old database.
Converting a CGI forum when millions of posts would not be a fun thing to do. Personally, I would just leave the old system in place and move on a new system for new posts. Just my 2 cents.
Thanks.
It helps to have those figures.
>> “There may come a time when the Perl system you have simply won’t cut it any more.” <<
.
Will you and I still be alive?
.
A rewrite is probably a formidable task, since this is an enterprise-level web app. My offhand guestimate is that it would take 2 senior developers, 1 fulltime tester, and one SOLID architect to pull off a full Free Republic implementation in 9 months to a year. Let us say, it is a year. Senior developers go about 100k, testers 60k, and an architect of the caliber you'd want is about 200k. That's 450k, and we haven't even talked about licensing development software and the equipment for them to develop on, much less the server farm itself. Figure 600k-750k to do a redevelopment.
This is true; but while Jim's stance seems somewhat anti-opensource of FR's internals I'm not sure if he is also unwilling to accept the 'donation' of developers re-implementing FR for him. I bet there's at least several capable programmer-types in the membership.
As I am currently unemployed [just graduated] I could probably donate some time to just such an effort. Even without pay, there is the possibility that I could put it on my resume as an experience point.
There may come a time when the Perl system you have simply won't cut it any more. When that happens, I suggest a conversion to a robust enterprise solution. I favor C#.NET with SQL back end. It's battle-tested and well-regarded for enterprise apps. John would be a pro in it in about a month.
As long as the system is being rewritten in another language, and the .NET paradigm of "language-independence" [CLR] is somewhat valid, then the choice of language should be given careful consideration. I would suggest that in this situation (where the common runtime makes data-types usable from anywhere) the best language would be one which facilitates: