On 14-Oct-06, at 10:09 PM, Saswata Banerjee & Associates wrote:
First, I am against the concept of cloning tally. Because tally is a very unsafe software. We refer to it as a "Time Bomb". I have explained this in the past on the group, but let me explain again.
In tally, you can insert an accounting transaction at any point of time behind in date. It will automatically renumber all vouchers and documents. There is no trail or any indication showing that it has been done. Similarly, you can delete an acccounting transaction at any point of time. Again, no one will be any wiser. In a corporate (or even SME Segment), this is dangerous as the accountant may manipulate the data for his own purpose, causing a loss to the organisation. I have made good money in the past by explaining this to the clients and sold them our services and moved them to alternate software.
agreed - we have discussed this before and it is my experience that tally is only used by accountants who are scrabbling around at the bottom of the food chain
Second, I think we should build web-based software. It is easier to run (everyone has broadband connection today), easier to maintain (you do not have to go to the clients office to solve the problem). It is also more popular platform. An added fact is that it works in case of multiple branch scenario and also allows owners to see the data from home.
as long as we separate presentation, business logic and data, there is no harm in having both web and gui interfaces. Note that high speed data entry is not possible on a web interface. We should plan for both
Third, with all respects to Kenneth, there are already existing accounting software that is good, but not designed for India. The 2 I like best is CKERP and WebERP. I have used both. Both have some faults and problems which can be solved. The advantage is that they are stable software and already used by people. You will need to add a few modules for taking care of Indian Tax Laws and providing for Indian GAP rules. It will be much faster than starting from scrap.
PHP will not cut it for a mission critical application - and if we are going the web+gui route - better to keep it in a language that can do both. Which rules out php anyway.
We may even strip down the software to remove things we dont plan to use (eg. CKERP has a small CRM module). and add things they dont have (Fixed Asset Register and automatic depreciation computation). Using this route, with a team of 5 programmers, in 3 months we will be ready to start shipping the software. You can give the software free and the Business Revenue Module will be by way of paid support and customisation. A 15 man-month project will not cost more than Rs. 250,000, I am sure funds can be raised even from this group.
no need to pay anyone - the bottleneck here is not programmers - it is domain knowledge
Abhisekh believes (I spoke to him on phone) that we will need to use AJAX so that we can give them keyboard hotkeys, etc. We can always do that in version 2. At least the software will be up, running and distributable in a minimum time frame.
AJAX and indian language support must be built in from the outset.
I remember Kennith had reservation against using php based software, primarily on security. But not being a programmer, I can not comment.
i *can* comment, but most of it would be unprintable ;-)
As for how people are going to make money from this - it is simple. No one wants/needs a pure FA package. We build a package which can be integrated into any custom package that the end-user has. And provide hooks for the retailer to add his custom modules. The retailer adds value by modifying - and others (including us) can make modules for sale/support. For example, the original version of Avsap was culled out of a bigger package which was written for a wholesaler in Amul products.